Experimental methods.

(A) The brief-access rig—multiple bottles, each made available periodically. To the right are example data schematizing the analysis of rhythmic licking to pull out lick numbers and “clusters,” both of which reflect palatability. (B) A schematic of in vivo GC recordings with taste delivery via a pre-implanted intra-oral cannula. (C) Histological verification of a representative electrode implant site (left) with dye marking the recording site in gustatory cortex. On the right is the schematic of the coronal slice through central GC, with the positions of electrode tips for all animals (N=9) denoted with red circles. (D) The experimental timeline, showing preference testing sessions and the delay (for surgery and recovery) between the final BAT day and Electrophysiology recording/passive tastant deliveries.

Individual and between-session differences in perceived taste palatability.

(A1-3) Data from first BAT sessions (B1-3) Data from final BAT sessions. (A1) Mean lick cluster sizes for two representative rats during the first BAT test illustrate clear individual differences in preference for sucrose, NaCl, and QHCl. A two-way ANOVA revealed a significant Taste × Animal interaction (F(2,65) = 13.44, p < 0.05). Post hoc Tukey tests further showed that sucrose, while highly preferred by the rat on the left, was perceived similarly to QHCl by the rat on the right (p < 0.05). (A2) Mean lick cluster sizes for each taste across all animals during the first BAT test (N = 9). The two rats shown in panel (A1) are outlined for reference. (A3) Differences in lick cluster sizes comparing sucrose and NaCl (top) and comparing citric acid (CA) and quinine (QHCl; bottom) during the first BAT test. One-sample t-tests (against a population mean of zero) revealed individual differences in the preferences for the two palatable and the two aversive tastes (p < 0.05). (B1-3) The same analyses as in A1-3, but for each rat’s final BAT session (depending on the rat, sessions 3 or 5). In (B3), we also show an analysis of whether individual rats changed their relative preferences between BAT sessions—two-way ANOVAs with variables Animal and Session, comparing the data shown in panels (A3) and (B3). The majority of rats exhibited significant changes in relative preference for tastes of the same valence across sessions (#, ps < 0.05). The labels R1–R9 denote the rats included in the data analysis.

Summary of taste stimuli used in brief-access tests (BATs) and electrophysiological recording sessions.

NaCl: sodium chloride; QHCl: quinine hydrochloride. The numbers listed after each taste in the Electrophysiology Sessions column indicate canonical rankings (4–3–2–1, from most palatable to most aversive).

Experience-related changes in taste preferences are not attenuation of neophobia.

Each panel shows the change in lick bout lengths between the first to final BAT session for either palatable (left) or aversive (right) tastes. In both cases, experience clearly changes behavior, but the changes aren’t in a consistent direction. See text for details.

Palatability-relatedness of Late-epoch taste responses match the individuals’ preferences.

(A) first- and final-session preferences (lick cluster size) for the battery of tastes sampled by one representative rat. The preference pattern changes with experience, and is non-canonical in the final session. (B) Peristimulus time histograms for the same tastes, passively delivered through an intra-oral cannula during an electrophysiology session, for a representative cortical neuron in the same rat, showing typical response dynamics including correlation (r-squared, dotted line and right y-axis) to palatability peaking around 1 second from taste delivery. The inset shows example waveforms. (C) Palatability correlations from the GC unit in (B) calculated based on the canonical ranking or the ranks estimated from the final BAT test. (D) The same analysis performed for the entire ensemble of neurons in this example rat.

The correlation between Late-Epoch GC firing and an individual rat’s taste preferences is specific to the most recent BAT assay.

(A). The palatability correlation over time from all of the (n=129) recorded neurons, using canonical palatability (green), lick cluster data from the first BAT session (blue), and lick cluster data from the final BAT session (orange). (B) The difference in correlation calculated from canonical palatability and data from the animal’s individual preferences (First BAT or Final BAT), separated into the Identity epoch (200-700ms) and the Palatability epoch (700-1200ms). * p < 0.01, *** < 0.001. (C) Magnitudes of the Late-Epoch palatability response (quantified in terms of the area under the curve in the correlation function shown in 5A) for 10,000 mild shuffles (reversing the order of nearest neighbors—see text for details) of rats’ preference orders. The vertical red dashed line shows the average magnitude for the real data.

A single session of IOC taste delivery to passive rats nullifies the correlation with palatability calculated from the most recent BAT assay.

(A) Cortical coding no longer matches rats’ individual preferences when using neurons (n=101) from electrophysiology session 2 (same conventions used in Figure 5A. (B) A summary of the data in (A), showing there is no longer a significant difference using preference rankings estimated from BAT performance. The Identity epoch and the Palatability epoch activity are taken from 200-700ms and 700-1200ms post-taste delivery, respectively.

Electrophysiological sessions do not change basic parameters of GC neural activity.

(A) Normalized activity levels of neurons recorded in the 2 electrophysiology sessions. The first session did not change neuron activity levels. (B) The percent of neurons showing significant taste responses is the same for the two sessions (session 1, 70.5% (91 units), session 2, 71.3% (72 units)). (C) Similarly, the proportion of neurons producing palatability-related responses in the two sessions did not differ (session 1, 35.7% (46 units), session 2, 29.3% (29 units)). (D) The absolute differences in session-1 and session-2 response magnitudes (a number that is necessarily positive) are similarly small—less than 0.6%—for all tastes.