Associative learning changes cross-modal representations in the gustatory cortex

  1. Roberto Vincis  Is a corresponding author
  2. Alfredo Fontanini  Is a corresponding author
  1. State University of New York at Stony Brook, United States
6 figures

Figures

Figure 1 with 1 supplement
Neural representation of different sensory modalities in GC of untrained rats.

(A) Percentage of neurons (n = 135) modulated (excited [gray] or inhibited [white]) by each stimulus. Odor and somatosensory (air puff) stimuli modulate a larger number of neurons compared to tone and light. Asterisks indicate post-hoc test corrected for multiple comparisons (Marascuillo’s test, p<0.05). (B) Pie chart showing the proportion of taste selective neurons (n = 87), that are modulated exclusively by taste (Taste only) or by taste and either one (Taste & 1) or two (Taste & 2) cross-modal stimuli. No neuron was modulated by taste and three or four cross-modal stimuli. (C) Pie chart showing the proportion of cross-modal responsive neuron (n = 45), modulated by one (1 Stimulus), two (2 Stimuli) or three (3 Stimuli) out of the four stimuli used. No neuron was modulated by all four the stimuli. (D) Raster plots and PSTHs of a representative GC neuron showing a broadly tuned response to the four tastants. (E) Raster plots and PSTHs of a GC neuron (same as the one showed in panel D) showing no response to cross-modal stimuli (air puff, odor, tone and light). (F) Raster plots and PSTHs of a second representative GC neuron showing a selective response to taste. (G) Raster plots and PSTHs of a GC neuron (same as the one showed in panel D) showing a selective response to a puff of air and no responses to Odor, Tone and Light. For D and F: vertical lines at time 0 indicate the onset of the stimulus. S for sucrose, N for NaCl, C for citric acid and Q for quinine. For E and G: grey shaded areas indicate the time and duration of cross-modal stimulus presentation. Red dotted line represents the time-course of orofacial movement activity.

https://doi.org/10.7554/eLife.16420.003
Figure 1—source data 1

Percentage of GC neurons responding to cross-modal and gustatory stimuli in untrained rats.

https://doi.org/10.7554/eLife.16420.004
Figure 1—figure supplement 1
Electrodes placement and experimental design for the group of untrained rats.

(A) Schematic representation of a coronal section of rat’s brain highlighting the dorso-ventral range of recording and electrodes placement in the gustatory cortex for the right (light blue) and left (light orange) hemisphere. (B) Schematic of the experimental design for each trial. Red bars highlight stimulus delivery. Each tastants is delivered through the IOCs in aliquots of 40 µl (duration of pulse: 25–30 ms) while each cross-modal stimuli is presented for 2 s. (C) Average mouth movements evoked by tastants (black) and by cross-modal stimuli (blue). Red dotted line indicates stimulus onset. Asterisk indicates p<0.05 Wilcoxon Test.

https://doi.org/10.7554/eLife.16420.005
Figure 2 with 1 supplement
Associative learning for different cross-modal stimuli.

(A) Left most panel: Average time course of orofacial movements evoked by cross-modal cues and sucrose during the first (gold) and 14th (cyan) day of classical conditioning (n = 5 rats). Shaded grey area and red dotted line indicate cues and sucrose presentation respectively. The shaded area around the curve represents ± SEM. Right most panel: average across trained rats (n = 5) of mouth movements (assessed by changes in pixel intensity) in different temporal epochs (1 s before cue onset [−1 to 0 s; Spont.], 1 s after cue onset [0.5–1.5 s; Cue] and 1 s after sucrose onset [2.7–3.7 s; Taste]) for the first (gold) and 14th (cyan) day of classical conditioning. (B) Average (n = 5 rats) of mouth movements across the 14 conditioning days. For A and B: asterisks indicate post-hoc tests corrected for multiple comparisons (Tukey, p<0.05). Rows from 1 to 4 show the orofacial activity evoked by air puff, odor, tone and light respectively.

https://doi.org/10.7554/eLife.16420.006
Figure 2—source data 1

Values of the orofacial movements evoked by the cross-modal stimuli.

https://doi.org/10.7554/eLife.16420.007
Figure 2—figure supplement 1
Experimental design for cue-taste association experiments and assessment of conditioned responses.

(A) Schematic of the experimental design used for cue-taste Pavlovian classical conditioning. Red bars highlight stimulus delivery. Between cue-offset and the delivery of sucrose there is a delay of 500 ms. Each cue is presented in a pseudo-random fashion and always precedes the delivery of sucrose. Conditioning lasted 14 days with and every day each rat received 20 trials of each cue-sucrose pair. See experimental procedures for further details. (B) Representative record of mouth movements, averaged across trials (Cue is odor; the trace pertain to the first day of training) for a single session (first day of classical conditioning). Note the different temporal epochs used for assessing the time-course of classical conditioning. Sucrose is delivered through IOCs in aliquots of 40 µl. Each cross modal stimulus is presented for 2 s. Between the cross modal stimuli offset and sucrose onset there is a constant delay of 500 ms.

https://doi.org/10.7554/eLife.16420.008
Figure 3 with 2 supplements
Neural representation of different sensory modalities after cue-taste association.

(A) Percentage of neurons (n = 118) modulated (excited [gray] or inhibited [white]) by each tastants and anticipatory cues. Asterisks indicate post-hoc test corrected for multiple comparisons (Marascuillo’s test, p<0.05). (B) Pie chart showing the proportion of taste selective neuron (n = 68) that are modulated exclusively by taste (Taste only) or by taste and one (Taste & 1), two (Taste & 2), three (Taste & 3) or four (Taste & 4) anticipatory cues. (C) Pie chart showing the proportion of cue responsive neuron (n = 69), which are only modulated by one (1 Cue), two (2 Cues), three (3 Cues) or all anticipatory cues (4 Cues). (D) Raster plot and PSTH of a representative GC neuron showing a significant excitatory response to sucrose. (E) Raster plot and PSTH of a GC neuron (same as the one showed in panel D) featuring significant excitatory responses to Air puff and Odor and no responses to Tone and Light. (F) Raster plot and PSTH of a GC neuron showing inhibitory responses to multiple tastants. (G) Raster plot and PSTH of a GC neuron (same as the one showed in panel F) displaying significant inhibitory responses to multiple cues (air puff, odor, tone and light). For D and F: vertical lines at time 0 indicate the onset of the taste delivery. S for sucrose, N for NaCl, C for citric acid and Q for quinine. For E and G: gray rectangular areas indicate the period of cue presentation. Vertical dotted lines at time 2.5 indicate the onset of sucrose delivery. Red dotted line represents the time-course of orofacial movement activity.

https://doi.org/10.7554/eLife.16420.009
Figure 3—source data 1

Percentage of GC neurons responding to cross-modal and gustatory stimuli in trained rats.

https://doi.org/10.7554/eLife.16420.010
Figure 3—figure supplement 1
Electrode placement and experimental design for the group of trained rats.

(A) Schematic representation of a coronal section of rat’s brain highlighting the dorso ventral range of recording and electrodes placement in the gustatory cortex for the right (light blue) and left (light orange) hemisphere. (B) Sketch of the experimental design for each trial. Red bars highlight stimulus delivery. See experimental procedures for further details. Each tastants is delivered through the IOCs in aliquots of 40 µl. Each cross modal stimulus is presented for 2 s. Between the cross modal stimuli offset and sucrose onset there is a constant delay of 500 ms.

https://doi.org/10.7554/eLife.16420.011
Figure 3—figure supplement 2
The neural bias for somatosensory and olfactory stimuli does not depend on the number of sessions at final performance level.

(A) Percentage of GC neurons modulated by cross-modal stimuli after matching the number of recording session at final performance level. (B) Distribution of the number of cross-modal neurons across recording days following reaching final performance level. Each full circle represents the number of responsive neurons recorded for each experimental day. Different colors represent responses to different cross-modal stimuli (see legend). Straight lines represent the best linear fit through each point.

https://doi.org/10.7554/eLife.16420.012
Comparison of cross-modal GC responsiveness between untrained and trained animals.

(A) Percentage of GC neurons that are modulated by cross-modal stimuli in untrained (blue) and trained (black) rats. (B) Percentage of taste selective GC neurons that are modulated only by tastants (Taste Only) or by tastants and at least one cross-modal stimulus (Taste & Cue); blue: untrained, black: trained. (C) Percentage of cue responsive GC neurons that are modulated only by one or more anticipatory cross-modal cues in untrained (blue) and in trained (black) rats. Asterisk indicates p<0.05 for test for equality of proportion.

https://doi.org/10.7554/eLife.16420.013
Time course of GC responses in the groups of trained and untrained rats.

(A) Population plot of all GC neurons modulated by at least one cross-modal stimulus in untrained (left panel) and in trained (right panel) rats. Each row represents a GC neuron. The color of each square along the x-axes (see color bar) represents the normalized firing rate within each 200 ms bin. The red dotted rectangular box indicates stimulus presentation. Neurons are clustered by the sensory modality (S for somatosensory, O for odor, T for tone and L for light) and ranked with excitatory and inhibitory responses from top to bottom respectively. (B) Histograms showing the percentage of cue responsive neurons that are excited and inhibited by anticipatory cues. (C) Histogram showing the average number of firing rate modulations in neurons responsive to cross-modal stimuli. Asterisk indicates p<0.05 for test for equality of proportion and t-test for panel B and C respectively.

https://doi.org/10.7554/eLife.16420.014
Figure 5—source data 1

Fraction and percentage of GC neurons excited or inhibited by the cross-modal cues.

https://doi.org/10.7554/eLife.16420.015
Coding of different sensory modalities for untrained and trained rats.

(A) Time course of the classification performance for cross-modal stimuli in the group of untrained rats. (B) Time course of the classification performance for cross-modal anticipatory cues in the group of rats that underwent an associative learning. For A and B: solid line represents correctly classified trials (as% ) based on population activity; shading gray area around traces represents 95% bootstrapped CI. Thick black horizontal lines below traces indicate significance from the chance level (95% bootstrapped CI above chance). Dotted horizontal lines indicate chance performance. Red dotted boxes represent the period of cross-modal stimulus presentation. (C) Left most panel: histograms for average cross-modal stimuli classification performance for neurons recorded in untrained rats. Classification performance is evaluated for a 200 ms time bin after stimulus onset. Error bars indicate SD. Right most panel: average classification for correct hits and false hits (second highest classification values after correct hits). Error bars indicate SEM. (D) Left most panel: as in C, but for a 200 ms time bin after stimulus offset. (E) Left most panels: histograms showing cross-modal cues classification performance for neurons recorded in trained rats. Classification performance is evaluated for a 200 ms time bin after stimulus onset. Error bars indicate SD. Right most panel: average classification for correct hits and false hits (second highest classification values after correct ones). Error bars indicate SEM. (F) Left most panels: as in E, but 200 ms after stimulus offset. (G) The time course of the CI measured for 200 ms bins over 1 s before and 2.5 s after cue onset in untrained rats. (H) Time course of the CI measured for 200 ms bins over 1 s before and 2.5 s after cue onset in trained rats. For C, D, E, F: each group of bars shows the percentage of trials classified for each stimulus. Labels under each bar indicate the actual delivered stimulus (S for somatosensory, O for odor, T for tone and L for light). For G and H: solid lines represent correct CI (measured assuming somatosensory and odor cues being similar). Shading around traces represents 95% bootstrapped CI (where the similarity between cues is shuffled, for example tone-odor or light-somatosensory being similar). Black horizontal lines indicate bins in which the CI is significantly different from chance. Asterisk indicates in panels C, D and E indicate p<0.05 for t-test.

https://doi.org/10.7554/eLife.16420.016
Figure 6—source data 1

Average and standard error of 'correct' and 'false' hits.

https://doi.org/10.7554/eLife.16420.017

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Roberto Vincis
  2. Alfredo Fontanini
(2016)
Associative learning changes cross-modal representations in the gustatory cortex
eLife 5:e16420.
https://doi.org/10.7554/eLife.16420