Distinct representations of basic taste qualities in human gustatory cortex

Abstract

The mammalian tongue contains gustatory receptors tuned to basic taste types, providing an evolutionarily old hedonic compass for what and what not to ingest. Although representation of these distinct taste types is a defining feature of primary gustatory cortex in other animals, their identification has remained elusive in humans, leaving the demarcation of human gustatory cortex unclear. Here we used distributed multivoxel activity patterns to identify regions with patterns of activity differentially sensitive to sweet, salty, bitter, and sour taste qualities. These were found in the insula and overlying operculum, with regions in the anterior and middle insula discriminating all tastes and representing their combinatorial coding. These findings replicated at super-high 7 T field strength using different compounds of sweet and bitter taste types, suggesting taste sensation specificity rather than chemical or receptor specificity. Our results provide evidence of the human gustatory cortex in the insula.

Fig. 1

Univariate analysis of voxel-specific taste tuning in the insula. a Upper two rows indicate activation maps for sour, sweet, bitter, salty, and tasteless stimuli of two subjects S1 and S2. The third row indicates group results when contrasted against baseline fMRI activity (n = 20 participants). The fourth row indicates group results when contrasted against tasteless, to control for non-taste-related gustatory activity and swallowing. When estimated against tasteless, controlling for positive and negative hedonic values, no taste-sensitive voxels survived statistical significance. A: anterior, P: posterior. b Each participant’s voxel activity in odd runs were aligned based on rank-ordered sensitivity to each sour, sweet, bitter, and salty taste in even runs, then averaged across participants. The corresponding downward trend for all tastes in each panel shows a lack of taste specificity. c Correlations were computed for voxel activation between odd and even runs between all taste combinations within each participant, submitted to one-sample t-test across participants. Corresponding activation between taste types shows a lack of taste-specific voxel tuning in the insula. d Correlations between odd and even runs for all same taste and different taste combinations within each participant. Correlation coefficients were z-transformed and subject to one-sample t-test across participants. Boxes represent the median and 25th/75th percentiles and whiskers represent the minimum and maximum. Bi: bitter, Dif: different, Sa: salty, So: sour, Sw: sweet. ✝p < 0.05 uncorrected, *p < 0.05 after Bonferroni correction for multiple comparisons (FWE < 5%)

Fig. 2

Multivoxel pattern analysis of taste-specific tuning in the insula. a Upper two rows indicate discriminability maps for sour, sweet, bitter, and salty stimuli of individual subjects (same as in Fig. 1a). Discriminability map indicates regions in which the average pairwise classification performance for a specific taste vs. each other taste was significantly higher than chance (50%). The linear discriminability searchlight maps (searchlight radius = 4 mm) were thresholded at p < 0.05 (uncorrected), which corresponded to 58.7% discrimination (chance level: 50%). The third row indicates group multi-taste-coding results (n = 20 participants) in which subjects were treated as a random effect, thresholded at FWE < 5% (small volume correction). A: anterior, P: posterior. b The number of taste types represented in the insula. Multi-taste population codes discriminated four basic tastes in the anterior and middle insula. c Overlap of four taste discriminability maps (each satisfies FWE < 5%) by leave-one-subject-out procedure. Color code indicates the percentage among the 20 leave-one-subject-out maps that was significant. d Representational dissimilarity matrix showing the pattern discriminability for each taste pair based on fMRI activity patterns of the right and left insula. e Representational dissimilarity matrix showing the pattern discriminability for each taste pair based on positive and negative valence ratings. f Classification performance for each taste pair compared with valence distance suggests discrimination of taste type separate from valence. Error bars indicate SE. g Repeated-measures ANOVA with taste type and valence as factors and trial fMRI activity pattern similarity (measured as correlations) as dependent variable. A main effect of taste type, but not valence nor interaction, demonstrates insular taste-type patterns do not reflect differences in experienced valence. Boxes represent the median and 25th/75th percentiles and whiskers represent the minimum and maximum. ✝p < 0.05 uncorrected, *p < 0.05 after Bonferroni correction for multiple comparisons (FWE < 5%)

Fig. 3

Multivoxel patterns supporting taste types outside the insula. Whole brain searchlight analysis revealed the frontal and parietal operculum also discriminate four taste types (n = 20 participants). FO: frontal operculum, PO: parietal operculum

Fig. 4

Specificity of taste quality not chemical compound in the insula under super-high field strength. a Correlations of voxel activation between odd and even runs between all taste combinations within each participant, submitted to one-sample t-test across participants (n = 11 participants). Statistically significant correlations within taste types but not across taste types suggest voxel-specific tuning at the individual level. b Correlations between odd and even runs for each same taste and different taste combinations in a show within-taste-type correspondence. Boxes represent the median and 25th/75th percentiles and whiskers represent the minimum and maximum. c The same correlations in a computed for voxels defined by group-level sensitivity to taste stimuli for each subject in a leave-one-out procedure. One-sample t-test across participants showed no taste specificity at the group level. d Correlations between odd and even runs for each same taste and different taste combinations in c reiterate lack of group-level taste specificity. Boxes represent the median and 25th/75th percentiles and whiskers represent the minimum and maximum. e Multivoxel pattern analysis shows group-level discriminability maps for sweet 1, sweet 2, bitter 1, and bitter 2 stimuli. f The number of taste stimuli represented in the insula. This replicates multi-taste population codes representing and discriminating multiple tastes in the anterior/middle insula as is shown in Fig. 2b. ✝p < 0.05 uncorrected, *p < 0.05 after Bonferroni correction for multiple comparisons (FWE < 5%)