Fig. 1: Distinct mechanisms underlie within-day and across-day habituation in A1.
From: Orbitofrontal cortex drives predictive filtering of sensory responses
a, Schematics illustrating two theories explaining sensory habituation by top-down predictive mechanisms. b, Representative two-photon image of A1 L2/3 neurons expressing GCaMP6s. c, Schematic illustrating the 5-day auditory habituation paradigm. d, Heatmaps of sound-evoked responses in neurons imaged across 5 days in a representative mouse. Neurons are sorted by their responses on day 1. e, Average (solid line) and s.e.m. (shading) of the Change index across responsive neurons for excitatory (red) and inhibitory (blue) sound-evoked responses over days. n = 879 responsive neurons from 20 mice. f, Change index of excitatory responses in individual neurons on day 5 compared to day 1 across all mice. Black line on the right represents mean. n = 879 responsive neurons. ****P = 1.8 × 10−91 (two-sided Wilcoxon signed-rank test). g, Change in sound-evoked response traces from day 1 to day 5 averaged across all significantly responsive cells (left). Black bar, 7-s sound. The amplitudes of the difference trace at 1 and 7 s after sound onset (right). ****P = 1.1 × 10−67 (two-sided Wilcoxon signed-rank test). Box, 25th to 75th percentiles; whiskers, 1 × interquartile range (IQR) (95.7% of data); black line, median; red line, mean; outliers not shown. h, Histograms of changes in response magnitudes in all neurons. Orange and green bars show neurons with significant increase and decrease. ****P = 4.3 × 10−138 (two-sided Fisher’s exact test). i, Schematic illustrating vectors representing within-day habituation (purple, Modewithin) and across-day habituation (green, Modeacross) within a high-dimensional space. Each dimension corresponds to the response magnitudes of individual neurons. j, Projection of trial-to-trial sound-evoked A1 ensemble activity dynamics onto the Modewithin (left) and Modeacross (right) vectors. Data points represent individual trials (100 trials × 5 days). Ensemble activity patterns repeated fast daily plasticity along the Modewithin axis, whereas they show continuous slow plasticity along the Modeacross axis across 5 days. n = 20 mice, 2,398 cells imaged throughout the 5 days. k, Cosine similarity between Modewithin and Modeacross, indicating a nearly orthogonal relationship. l, Left, change in sound-evoked ensemble activity along Modewithin from trial 1 to trial 10, demonstrating prominent within-day, across-trial habituation around the tone onset. Right: Change in sound-evoked ensemble activity along Modeacross from day 1 to day 5. Across-day habituation slowly ramps up during the sustained tone. m, Summary data showing the Ramp-up Index for across-day plasticity (day 1 to day 5) and within-day plasticity (trial 1 to trial 10). n, Top left, schematic illustrating pupil monitoring during two-photon calcium imaging. Top right: Representative image from a pupil camera. Bottom right, representative pupil diameter dynamics during sound presentations. Black lines indicate tone timings. o, Change in normalized pupil diameter across days. n = 17 mice. p, Two alternative models illustrating the dependence of neuronal sound responses on normalized pupil diameter. Left, a model in which neuronal habituation depends on the decrease in arousal level. Right, a model in which neuronal habituation is independent of arousal level. Individual dots represent trials; black, day 1 trials; red, day 5 trials. Lines indicate regression lines. q, Experimental data showing neuronal response magnitudes binned by normalized pupil diameter separately for day 1 (black) and day 5 (red). Day 5 neuronal responses are smaller than day 1 responses regardless of normalized pupil diameter. n = 10 mice, 977 cells. Day 1 versus day 5, ****P = 2.0 × 10−30 (unbalanced two-way analysis of variance). Error bars and shading represent mean ± s.e.m.
