Frontal cortex activation causes rapid plasticity of auditory cortical processing

Abstract

Neurons in the primary auditory cortex (A1) can show rapid changes in receptive fields when animals are engaged in sound detection and discrimination tasks. The source of a signal to A1 that triggers these changes is suspected to be in frontal cortical areas. How or whether activity in frontal areas can influence activity and sensory processing in A1 and the detailed changes occurring in A1 on the level of single neurons and in neuronal populations remain uncertain. Using electrophysiological techniques in mice, we found that pairing orbitofrontal cortex (OFC) stimulation with sound stimuli caused rapid changes in the sound-driven activity within A1 that are largely mediated by noncholinergic mechanisms. By integrating in vivo two-photon Ca(2+) imaging of A1 with OFC stimulation, we found that pairing OFC activity with sounds caused dynamic and selective changes in sensory responses of neural populations in A1. Further, analysis of changes in signal and noise correlation after OFC pairing revealed improvement in neural population-based discrimination performance within A1. This improvement was frequency specific and dependent on correlation changes. These OFC-induced influences on auditory responses resemble behavior-induced influences on auditory responses and demonstrate that OFC activity could underlie the coordination of rapid, dynamic changes in A1 to dynamic sensory environments.

Figure 1.

Response changes in A1 depend on OFC activation. A, Left, Bright-field image of a brain slice in the frontal cortex of the mouse showing the tracks of the bipolar stimulating electrodes used for OFC activation. Boundaries of OFC are indicated with a drawing from mouse brain atlas. Right, Mapping of stimulation locations in electrophysiological experiments. B, Time-frequency analysis of ERPs at the PF before (left) and after (right) OFC-sound pairing episodes. R-spectra were computed for each responsive channel then averaged (see Materials and Methods). C, Same as in B and C except that OFC was silenced with a mixture of GABA agonists before the OFC-sound pairing episodes. No change in power was observed. D, Bright-field and fluorescence images showing locations of stimulation electrode and spread of muscimol in OFC. GABA agonist spread was tracked by rhodamine in the inactivating solution. The inactivated region encompasses the location of the stimulation electrode. E, Same as in B and C except that ACSF was injected into the OFC before the OFC-sound pairing episodes. F, G, Comparison of average induced gamma (30–80 Hz) power during the entire stimulus period (400 ms) before (Pre) and after (Post) OFC-sound pairing episodes for each responsive channel under control conditions (F) or after application of GABA agonists or saline. H, Desynchronization index. I, Difference in LFP peak amplitude tuning after pairing and with application of GABA agonists (“silenced”).

Figure 2.

OFC stimulation-induced response changes in A1 are not mediated by cholinergic mechanisms. A, B, Comparison of average induced gamma (30–80 Hz) power during the entire stimulus period (400 ms) before (Pre) and after (Post) OFC-sound pairing episodes for each responsive channel after A1 application of atropine (A) or mecamylamine (B). C, D, Comparison of desynchronization indices (C) and LFP peak tuning change (D) across different pharmacological conditions (atropine application in A1, mecamylamine application in A1).

Figure 3.

Plasticity of A1 neuron BF with OFC microstimulation and sound pairing. A, Left, Anatomical identification of imaging location in A1. Right, Neurons are loaded with Fluo4-AM and Ca²⁺ responses are measured with two-photon imaging. B, Fluorescence responses of selected neurons and tuning curves. C, Left, BFs of neurons in a single field of view before pairing (Baseline). Neurons denoted by circles; color: BF of neuron (see Materials and Methods). Color bar: frequency range. Right, BFs of the same cells after pairing with 23 kHz (Post-pairing). Scale bars: A, C, 25 μm.

Figure 4.

Plasticity of A1 neuron receptive fields with OFC microstimulation and sound pairing. A, Effects of OFC pairing paradigm on three representative neurons from the field in Figure 3C). Each column represents data from a single neuron. Numbers are indicated in the field in Figure 3D). Abscissa: Frequency. Rows 1 and 2 represent smoothed Ca²⁺ responses before OFC pairing (row 1) and after OFC pairing (row 2). Color bar: ΔF/F (±1%). Ordinate: Time from stimulus onset up to nine imaging frames later (∼1 s). Asterisks: BF before and after OFC pairing. Row 3: Frequency tuning curves derived from maximum ΔF/F during the stimulus period. Black, before OFC pairing (Baseline); Red, after OFC pairing; shading, 95% confidence intervals (see Materials and Methods). Ordinate: Max ΔF/F. Row 4: Overall change in response (Post-Baseline) of the significant (p < 0.05) ΔF/F responses for each neuron. For all rows: Dashed vertical line indicates PF. Row 5: Mean change over time showing the profile of spectral changes with 95% confidence interval. B, Example time-frequency response differences for seven individual neurons indicated in Fig. 3C). Axes as in (A) row 4 (1,4,and 7 same, now as contour plots). C, Mean change in response for all neurons in the field of view in 3D; color axis ±0.5%). D, Average DSTRMs after OFC pairing for 8 of 17 fields of view. Each plot shows the average difference (post-pre) in the responses of the neurons for each field of view. All DSTRMs are aligned to PF (dashed line). Scale is the same for all subplots as in C.

Figure 5.

Overall effect of OFC microstimulation and sound pairing in A1. A, Mean change in response for the population of imaged neurons that underwent OFC pairing. Color axis ±0.1%. B, Distribution of BFs of all neurons before (black) and after (gray) OFC pairing. Before pairing distribution shows that PFs were chosen such that different spectral distances (PF-BF) over ±3 octaves were represented in the population. The two peaks flanking PF (red line) indicate a tendency of PF to be on the flanks of the tuning curves rather than at BF of the neurons. C, Difference in numbers of neurons with particular BF before and after OFC Pairing. Fraction of neurons representing PF increases whereas the fraction of neurons representing other sound frequencies decreases. D, Mean change in response for the sound alone control. E, Population average of changes in STRMs for all neurons with a random PF selected for each imaged field (see main text and Materials and Methods for details on calculation). Axes as in A. F, Mean change over time showing spectral profile of changes in each case. The effective change in spectral profile due to the OFC pairing is shown in black after removing the effect of habituation caused by the presentation of a sound alone. Scale bar, 0.2%.

Figure 6.

Effect OFC microstimulation and sound pairing in A1 depends on stimulation strength. A, Mean changes in STRMs with OFC pairing in the population of neurons at different stimulation strengths. Strongest enhancements were observed at PF at the highest stimulation strength and weak changes at the lowest stimulations strength. Axes as in Figure 3A and B (color bar ±0.2%). B, Measures of response change plotted as a function of OFC stimulation strength, straight lines show the significant (**p < 0.05) best linear fits. All measures of magnitude of response change were positively correlated with OFC stimulation strength (Fig. 4).

Figure 7.

Two populations of response changes after OFC-sound pairing. A, Distributions of correlation values of actual response differences (after-before pairing) for each neuron with the mean effect shown in Figure 3A (black line) and of shuffled version of the response differences of same neurons (gray). Solid and dashed vertical black lines: mean of the distributions of actual and shuffled datasets; red and blue lines: ±2*STD of shuffled data. Performing the same calculation but computing the correlation value of each neuron's response change with the average of all other neurons' response change produced similar results. Inset shows the increased number of neurons with correlations <2 STD. B, Top, Average differences of the anticorrelated and positively correlated neurons (color axis ±0.4%). Bottom, Time-averaged mean spectral profiles of changes. C, Relationship of fraction of anticorrelated and positively correlated cells in each field as a function of stimulations strength. D, Mean response change in 0.5 octave bins around PF plotted as a function of spectral separation of before-BF from PF as changes are approximately symmetrical changes above and below PF. E, Mean ΔF/F change at PF and adjacent frequency bins as function of before-BF-PF for OFC pairing and sound alone. Dashed lines indicate regression lines. The slopes are significantly different (p < 0.05). Note that with sound alone no strengthening is observed.

Figure 8.

Time-lapse data show no specific changes. A, Cumulative distribution of correlation values of time-lapse response changes with the mean effect shown in Figure 3A. Distributions were generated by selecting a random PF for each imaged field (repeating 100 times for each imaged field) and aligning response changes according to the PF, then averaging. Response changes were also shuffled in frequency and time to create a shuffled distribution. Cumulative distribution functions are shown for correlation values with mean effect of actual differences (black) and shuffled differences (gray). This shows a lack of structure related to PF in the time-lapse data, with changes resembling random noise (p > 0.44, Kolmogorov–Smirnov test). B, Time-lapse changes in BF of A1 neurons. Distribution of changes in best frequency (BFBefore-BFAfter) in two sequential imaging sessions separated by ∼10 min. Fifty percent of neurons exhibit stable BF (within 0.75 octaves; dashed lines). The remaining 50% of imaged neurons had altered BF. With single-unit recordings in awake ferrets such changes measured for only linear cells with good spectrotemporal receptive fields (STRFs) find >25% of cells are labile (Elhilali et al., 2007).

Figure 9.

Changes in pairwise correlations between neurons produced by OFC pairing and sound alone. A, Distribution of pairwise signal correlation values between all simultaneously imaged pairs before OFC pairing (black). Distribution of correlation values of pairwise signal correlation values of frequency and time shuffled STRMs (gray). B, Top, Distributions of changes in pairwise signal cross-correlations between all simultaneously imaged pairs of neurons in A1. Red, OFC pairing condition; blue, sound alone condition. Overlap between the distributions is shown in purple. Cumulative distribution functions are overlaid (red, OFC pairing; blue: sound alone). Bottom, Difference between the two distributions on the top part (Pairing − sound alone). Black line indicates raw difference; red line indicates smoothed version of the difference. C, Fractional change in mean correlations before and after OFC pairing show a significant decorrelation relative to zero (p < 10⁻⁶) and the time-lapse case (p < 0.01) while that for sound alone showed a significant increase in mean correlations relative to zero (p < 0.01) and the time-lapse case (p < 0.001). D, Distribution of pairwise noise correlation values between all simultaneously imaged pairs before OFC pairing (black). Distribution of correlation values of pairwise noise correlation values of frequency and time-shuffled STRMs (gray). E, F, Conventions as in B and C, showing equivalent plots for noise correlations between pairs of neurons. Fractional change in mean noise correlations before and after OFC pairing shows a significant increase relative to zero (p < 0.05) and time-lapse case (p < 10⁻¹⁰) while that for sound alone showed a significant decrease from 0 (p < 10⁻⁴) and from the time-lapse case (p < 0.05).

Figure 10.

Discrimination of PF by single neurons. A, Distribution of mean discrimination performance of PF from each other frequency (PF/O) by all neurons before OFC pairing (black). The same distribution based on the calculation done on baseline activity (before stimulus onset; gray) shows minimal bias in the estimation (p < 10⁻⁹¹). B, Same distributions as in A plotted with sound alone and time-lapse data showing our choice of PF in the two cases and a random PF for the time-lapse data show no bias. C, Scatterplot showing discrimination performance by single neurons (n = 307) after OFC pairing as a function of discrimination performance before pairing. D, Bar graph showing median single neuron discrimination performance. Error bars indicate interquartile range. All groups are statistically the same (p > 0.1).

Figure 11.

Discrimination performance as a function of population size. A, Median performance by populations of cells as a function of population size (N = 1–14) in discriminating PF from other frequencies before (blue lines) and after (red lines) OFC pairing (left) and sound presentation alone (right). Error bars indicate interquartile ranges. Data show no bias. B, C, Actual change (Post-Pre, left) (B) and relative change (C) ([Post-Pre]/Pre, right) in discrimination. Changes at the largest population size were always significantly different (p < 0.05). D, Normalized discrimination (normalized to N = 1) as a function of population size is shown for pairing and sound alone as in A, black line indicates discrimination by equivalent populations assuming independence between cells. E, Relative changes in discrimination due to correlations as a function of population size (changes are significant for N = 14, p < 0.05) over the independent case.

Figure 12.

OFC sound pairing produces distinct types of response changes in A1 neurons. A, Fraction of variability of the observed STRM changes with OFC pairing explained by each PC and cumulatively starting from the strongest component shows that the first three PCs explain 70% of the variance, while each of the remaining explain <5% of the variance. B, Patterns of underlying response changes as defined by the first three PCs. The mean spectral profile of each PC is shown below. C, Absolute PC scores or distance of response change along each PC axis depends on OFC stimulation strength (p < 0.0001, 0.001, and 0.05, respectively, left to right).