Slide 1

Slide 1

Abuzar Mahmood

Katz Lab

The Cortico- Amygdalar Interaction Dynamics underlying Taste Perception & Action

![A picture containing dress, holding, standing, white

Description automatically generated](./media/ppt/media/image3.png “Picture 6”)

![A rodent looking at the camera

Description automatically generated](./media/ppt/media/image4.png “Picture 8”)

Illustrations courtesy of Katie Kimbrell

Slide 2

![A close up of an animal

Description automatically generated](./media/ppt/media/image5.png “Google Shape;308;p8”)

Cortex

Amy

Recent and Ongoing Projects

Slide 3

Rosenstein & Oster , 1988

Grossman et al, 2008

The Inevitable Sensorimotor Transformation of Taste Information

Sweet or salty – (yum!)

Vs

Bitter or sour –  (yuck!)

![A collage of a rat in a cage

AI-generated content may be incorrect.](./media/ppt/media/image8.png “Picture 3”)

![A collage of a rat in a cage

AI-generated content may be incorrect.](./media/ppt/media/image8.png “Picture 4”)

![A drawing of a cat

AI-generated content may be incorrect.](./media/ppt/media/image9.png “Picture 6”)

![A drawing of a bug\'s face

AI-generated content may be incorrect.](./media/ppt/media/image10.png “Picture 8”)

Start around 750-1250ms post-stimulus delivery (with some variability)

Lick

Gape

C:\baby pics\July-Sept 09\gaping with lemon.jpg

Slide 4

Recurrent Connectivity in the Taste Circuit

![A close up of an animal

Description automatically generated](./media/ppt/media/image12.png “Google Shape;128;p3”)

GC

Sensory Signals from Cranial Nerves

Carleton et al. 2010

Merlo et al. 2015​

Yang et al. 2025

Yan

Brainstem

Thalamus

Limbic Regions

Slide 5

Recurrent Connectivity in the Taste Circuit

![A close up of an animal

Description automatically generated](./media/ppt/media/image12.png “Google Shape;128;p3”)

GC

Sensory Signals from Cranial Nerves

Carleton et al. 2010

Merlo et al. 2015​

Yang et al. 2025

Slide 6

Background

GC neural response dynamics

Dynamics linked to onset of behavior

GC-Amygdala link

Results

GC-Amygdala dynamics are tightly linked

  • And yet...asymmetric
  • Asymmetric influences demarcate response states
  • Neurons most deeply embedded in circuit show strongest encoding

Preliminary

Intra-state dynamics

Outline

Slide 7

Suc

NaCl

CA

Quin

Adapted from Flores et al. 2018

Experimental Setup:

Data acquisition

Slide 8

![Chart, histogram

Description automatically generated](./media/ppt/media/image18.png “Google Shape;195;p42”)

-500 -250 0 250 500 750 1000 1250 —— —— — —– —– —– —— ——

5 — 4 3 2 1 0

Time post-stimulus delivery (ms)

Firing Rate (Hz)

Unpalatable

Palatable

GC

YUM

YUCK

Katz et al. 2001 | Jones et al. 2007 | Mahmood et al. 2023 ​

GC taste response dynamics

Slide 9

Quinine

1 sec

10

9

8

7

6

5

4

3

2

1

Simultaneously recorded neurons

Katz et al. 2001 | Jones et al. 2007 | Mahmood et al. 2023 ​

![Chart, histogram

Description automatically generated](./media/ppt/media/image18.png “Google Shape;195;p42”)

-500 -250 0 250 500 750 1000 1250 —— —— — —– —– —– —— ——

5 — 4 3 2 1 0

Time post-stimulus delivery (ms)

Firing Rate (Hz)

Unpalatable

Palatable

GC

YUM

YUCK

Deeper dive:

What happens on single trials?

Slide 10

Variable GC Dynamics across trials

Sadacca et al . 2016

Slide 11

Sadacca et al , 2016, Mahmood*  Baas-Thomas* et al. 2026

Ensemble state transitions reflect taste decision-making

Gaping Rat [IMAGE] | EurekAlert! Science News Releases

Palatability state

Gape

Also true for ingestive behaviors

Slide 12

Katz et al. 2001 | Jones et al. 2007 | Mahmood et al. 2023 ​

![Chart, histogram

Description automatically generated](./media/ppt/media/image23.png “Google Shape;201;p42”)

BLA

-500 -250 0 250 500 750 1000 1250 —— —— — —– —– —– —— ——

2 —– 1.5 1 0.5

Time post-stimulus delivery (ms)

Firing Rate (Hz)

Palatable

YUM

YUCK

BLA Taste Response Dynamics

Unpalatable

Slide 13

Lin et al, 2021

BLA input is needed for emotional processing in GC

Intact

GC

![Chart

Description automatically generated](./media/ppt/media/image26.jpg “Picture 6”)

Change in # of significant neurons

![A collage of graphs and charts

AI-generated content may be incorrect.](./media/ppt/media/image27.jpg “Picture 26”)

BLA

Slide 14

Background

GC neural response dynamics

Dynamics linked to onset of behavior

GC-Amygdala link

Results

BLA-GC interaction dynamics

Outline

Slide 15

GC is carrying out taste processing in concert with other regions

BLA and GC dynamics, on-average, appear to be coordinated

BLA input to GC is important for palatability processing

Need a direct test of BLA-GC coordination and characterization of directional influence

Where does this leave us?

Slide 16

Predictions - Part 1

BLA population activity will evolve as a sequence of states

BLA-GC transition #2 will be coordinated

Slide 17

Mahmood et al. 2023

Stim-aligned

Transition-aligned

BLA

States in BLA Population Activity

Slide 18

BLA population activity will evolve as a sequence of states

BLA-GC transition #2 will be coordinated

Predictions - Part 1

Slide 19

Mahmood et al. 2023

Amygdala and GC couple into a unit to process palatability

GC

BLA

1

2

3

transition

% of datasets with high BLA-GC correlations

GC

BLA

GC

BLA

GC

BLA

Slide 20

Mahmood et al. 2023

Amygdala and GC couple into a unit to process palatability

1

2

3

transition

% of datasets with high BLA-GC correlations

![A diagram of a transition

AI-generated content may be incorrect.](./media/ppt/media/image37.png “Picture 1”)

Slide 21

Amygdala and GC couple into a unit to process palatability

1

2

3

transition

% of datasets with high BLA-GC correlations

![A graph of different levels

Description automatically generated](./media/ppt/media/image38.png “Content Placeholder 4”)

Transition 1

GC

leads

BLA

leads

msec

![A graph of different levels

Description automatically generated](./media/ppt/media/image38.png “Content Placeholder 4”)

Transition 2

Transition 3

msec

msec

Mahmood et al. 2026

Slide 22

What's next for BLA and GC?

Caveats of transition coordination analysis

Snapshots of interaction

Symmetric

Slide 23

BLA will be driving GC to transition into palatability state (driving transition #2), after which GC emits behavioral command *

Neurons involved in inter-region communication will have stronger taste encoding.

*

Lin et al. 2021

Bechara, et al. 1999

Schoeanbaum, et al. 1998

Predictions - Part 2

Slide 24

BLA influences GC on the way to amygdala-cortical coupling

Granger Causality

Region A

Region B

Signal

Amplitude

Time

C B * B history = B future

C A * A history +

Slide 25

![A comparison of colored squares

AI-generated content may be incorrect.](./media/ppt/media/image39.jpg “Picture 16”)

![A comparison of colored squares

AI-generated content may be incorrect.](./media/ppt/media/image39.jpg “Picture 6”)

BLA influences GC on the way to amygdala-cortical coupling

Spectral Granger Causality

Frequency (Hz)

![A comparison of colored squares

AI-generated content may be incorrect.](./media/ppt/media/image39.jpg “Picture 5”)

Mahmood et al. 2026

-500 0 500 1000 1500

Slide 26

BLA will be driving GC to transition into palatability state (driving transition #2), after which GC emits behavioral command

Neurons involved in inter-region communication will have stronger taste encoding.

Predictions - Part 2

Slide 27

Input

History

History of

other neurons

History of

current neuron

Future of

current neuron

Embedding of neurons in the BLA-GC interactions is related to “tastiness”

Poisson Generalized Linear Modeling

Generalized Linear Model - an overview | ScienceDirect Topics

Generalized Linear Model - an overview | ScienceDirect Topics

Generalized Linear Model - an overview | ScienceDirect Topics

Generalized Linear Model - an overview | ScienceDirect Topics

Generalized Linear Model - an overview | ScienceDirect Topics

Slide 28

Embedding of neurons in the BLA-GC interactions is related to “tastiness”

Output

to BLA

20%

12%

17%

Taste

Specificity

Palatability

GC neurons

![A group of bar graphs

Description automatically generated](./media/ppt/media/image48.png “Content Placeholder 13”)

Mean discriminability

![A group of bar graphs

Description automatically generated](./media/ppt/media/image48.png “Content Placeholder 13”)

Influenced

by BLA

Mean palatability (r)

Mahmood et al. 2026

Slide 29

Obligatory summary slide

GC taste responses reflect taste processing across successive epochs.

Ensemble analysis of single trials reveals coding to emerge suddenly across ensemble transition.

These transitions in turn drives behavior… probably by modulating a brainstem CPG .

The mechanism of construction of this transition involves coupling of GC and BLA into a functional unit.

System functions under spatial “hierarchies”

Slide 30

Background

GC neural response dynamics

Dynamics linked to onset of behavior

GC-Amygdala Link

Results

Preliminary

Intra-state dynamics

Calia-Bogan, Katz, and Mahmood​ (In Prep)​

Outline

Slide 31

Incomplete list of papers from the lab using HMMs / changepoints for state transitions:

Jones, Lauren M., et al. “Natural Stimuli Evoke Dynamic Sequences of States in Sensory Cortical Ensembles.” Proceedings of the National Academy of Sciences, vol. 104, no. 47, Nov. 2007 , pp. 18772–77. DOI.org ( Crossref ), https://doi.org/10.1073/pnas.0705546104 .

Sadacca , Brian F., et al. “Sodium Concentration Coding Gives Way to Evaluative Coding in Cortex and Amygdala.” Journal of Neuroscience, vol. 32, no. 29, July 2012 , pp. 9999–10011. www.jneurosci.org, https://doi.org/10.1523/JNEUROSCI.6059-11.2012 .

Moran, Anan, and Donald B. Katz. “Sensory Cortical Population Dynamics Uniquely Track Behavior across Learning and Extinction.” The Journal of Neuroscience, vol. 34, no. 4, Jan. 2014 , pp. 1248–57. PubMed Central, https://doi.org/10.1523/JNEUROSCI.3331-13.2014 .

Sadacca , Brian F., et al. “The Behavioral Relevance of Cortical Neural Ensemble Responses Emerges Suddenly.” Journal of Neuroscience, vol. 36, no. 3, Jan. 2016 , pp. 655–69. www.jneurosci.org, https://doi.org/10.1523/JNEUROSCI.2265-15.2016 .

Mukherjee, Narendra, et al. “Impact of Precisely-Timed Inhibition of Gustatory Cortex on Taste Behavior Depends on Single-Trial Ensemble Dynamics.” eLife , edited by Laura L Colgin et al., vol. 8, June 2019 , p. e45968. eLife , https://doi.org/10.7554/eLife.45968 .

Lin, Jian-You, et al. “Perturbation of Amygdala-Cortical Projections Reduces Ensemble Coherence of Palatability Coding in Gustatory Cortex.” eLife , vol. 10, p. E65766. 2021 PubMed Central, https://doi.org/10.7554/eLife.65766. Accessed 25 Mar. 2025.

Mahmood, Abuzar, et al. “Coupled Dynamics of Stimulus-Evoked Gustatory Cortical and Basolateral Amygdalar Activity.” Journal of Neuroscience, vol. 43, no. 3, Jan. 2023 , pp. 386–404. www.jneurosci.org, https://doi.org/10.1523/JNEUROSCI.1412-22.2022 .

Svedberg, Daniel A., and Donald B. Katz. Neural Correlates of Rapid Familiarization to Novel Taste. bioRxiv , 9 May 2024 . bioRxiv , https://doi.org/10.1101/2024.05.08.593234 .

Baas-Thomas, Natasha, et al. “The Ingestive Response Reflects Neural Dynamics in Gustatory Cortex.” 2 Oct. 2025. Neuroscience, https://doi.org/10.1101/2025.10.01.679845.

Mahmood, Abuzar, et al. “Sensory and Palatability Coding of Taste Stimuli in Cortex Involves Dynamic and Asymmetric Cortico- Amygdalar Interactions.” 4 July 2025. Neuroscience, https://doi.org/10.1101/2025.07.01.662567 .

States in GC assuming uniform emissions

Slide 32

Stochastic Transitions between Neural States in Taste Processing and Decision-Making | Journal of Neuroscience

Egest

Ingest

Miller and Katz 2010; 2013, Mazzucato et al. 2015

I’d recommend not motivating this simply with “when you look closer…”; I’d go with something more theory-driven, about how it makes sense that there should be internal dynamics, and also talk about the problem of explaining why the state transitions happen when they happen (which you can then expand upon when you get to the slide about external vs internal driving).

Attractor dynamics

Slide 33

Potential Intra-state Dynamics

Slide 34

Adapted from Jones et al. 2007

Vincent Calia-Bogan

If we do look closer…

Slide 35

Intra-state dynamics underlie state transitions and decision-making in go/no go tasks in mice

Adapted from Luo et al. 2023. See also Lindermann et al. 2020 and Escola et al. 2011

Existing literature: various models with intra-state dynamics fit quite well

Gap: None of these models actually characterize how the state dynamics are related to neural processing.

Goal: link back to neural mechanisms underlying state transitions.

Z0

Z1

Current Literature

Slide 36

Competing mechanisms underlying intra-state dynamics

H1 : Observed state transitions are externally triggered

GC intra-state dynamics are “clipped” differently on different trials

H2 : Internal dynamics trigger state transitions

Intra-state dynamics reach an “end-point”

Adapted from Carlton et al. 2010 and Merlo et al. 2015

Internally vs Externally-driven Dynamics

Slide 37

Autoencoder  --> N onlinear dimensionality reduction

LSTM  --> temporal relationship

Population Dynamics: Autoencoder RNN

Slide 38

Trial 9

Trial 10

Lat 1

Lat 2

Lat 4

Lat 3

HMM <-> RNN Convergent results

Slide 39

Smooth, non-uniform dynamics

Alpha (8-10Hz) oscillations

Stay tuned for testing of mechanistic hypotheses

Lat 1

Lat 2

Lat 4

Lat 3

Lat 1

Lat 2

Lat 4

Lat 3

Two primary types of dynamics

Slide 40

Obligatory summary slide

GC taste responses reflect taste processing across successive epochs.

Ensemble analysis of single trials reveals coding to emerge suddenly across ensemble transition.

These transitions in turn drives behavior… probably by modulating a brainstem CPG .

The mechanism of construction of this transition involves coupling of GC and BLA into a functional unit.

System functions under spatial and temporal “hierarchies”

Slide 41

![A diagram of a brain

AI-generated content may be incorrect.](./media/ppt/media/image62.png “Google Shape;393;p30”)

Amygdala as a modulator for learning rate / encoding volatility

Changes in Amygdala response / GC-BLA interaction through the session

GC-BLA-Thalamus interaction

Explanation in terms of Global Workspace Theory

Redundant vs. synergistic processing

Future directions

Slide 42

Taste-Dopamine axis questions / thoughts

Single-trial link between onset of behavior in passive taste paradigm and dopamine ramps?

Effect of dopamine ramps on behavior is dependent on inter-region connections / network context?

That’s why activating VTA by itself does not generate behavior?

Existence of vectorized / contextualized RPE in VTA suggests top-down modulated (Langdon and Schoenbaum)

What influences? Any feedback from reward circuit to sensory?

  • Competition between learned task and urgency for consumption occurs at what level?

Slide 43

![A group of people posing for a photo

AI-generated content may be incorrect.](./media/ppt/media/image66.png “Picture 2”)

  • The Katz
  • The Katz lab
  • Jessica Steindler
  • Hannah Germaine
  • Vincent Calia-Bogan

Acknowledgements ​

![A person wearing glasses and a blue shirt

AI-generated content may be incorrect.](./media/ppt/media/image67.jpg “Picture 2”)

Rattus Norvegicus

Slide 44

Thank you for your attention.

GC taste response dynamics

![Chart, histogram

Description automatically generated](./media/ppt/media/image18.png “Google Shape;195;p42”)

-500 -250 0 250 500 750 1000 1250 —— —— — —– —– —– —— ——

5 — 4 3 2 1 0

Time post-stimulus delivery (ms)

Firing Rate (Hz)

Unpalatable

Palatable

GC

YUM

YUCK

Katz et al. 2001 | Jones et al. 2007 | Mahmood et al. 2023 ​

Population Dynamics: PCA

PCA limitations:

No temporal dependencies (as part of model)

Only linear relationships

Single trial

Supplemental: RNN works well

Comparison of binned firing rates and RNN-inferred firing rates:

Time (bins)

Firing rate

Time (bins)

Time (bins)

Unit 1

Unit 2

Unit 3

Unit 2

RNN

Binned

Population dynamics: cleaning via PCA

RNN Single trial

Slide 49

Effect of BLA perturbation on GC processing

![A picture containing timeline

Description automatically generated](./media/ppt/media/image81.jpg “Picture 2”)

  • Bechara, A., et al. “ Different Contributions of the Human Amygdala and Ventromedial Prefrontal Cortex to Decision-Making. “ 1999, https://doi.org/10.1523/JNEUROSCI.19-13-05473.1999.
  • Schoenbaum, G., et al. “ Neural Encoding in Orbitofrontal Cortex and Basolateral Amygdala during Olfactory Discrimination Learning .” 1999, https://doi.org/10.1523/JNEUROSCI.19-05-01876.1999.

Slide 50

Alignment of ingestive behaviors and GC dynamics

Slide 51

![A screenshot of a graph

AI-generated content may be incorrect.](./media/ppt/media/image83.jpg “Picture 1”)

GLM Fit Quality Assessment

Slide 52

STATES IN BLA POPULATION ACTIVITY

Mahmood et al. 2023

Destroy population coordination

Destroy sharp changes

Trial 1

Trial 2

Slide 53

Samuelsen 2013

Slide 54

Saravani 2019, epoch

Slide 55

Forseth 2021 – distributed interaction

Slide 56

Global Workspace Theory

(Gaillard 2009)

Bouaichi 2020, Licking-related GC Dynamics

![A diagram of a mouse

Description automatically generated](./media/ppt/media/image92.jpg “Picture 6”)

![A diagram of a mouse

Description automatically generated](./media/ppt/media/image92.jpg “Picture 7”)

Slide 58

Mahmood et al, 2022

GC

BLA

Ctx

Amyg

Ctx

Amyg

Ctx

Amyg

Amygdala and GC become a unit to process palatability

GC

BLA

Slide 59

0.25

Gustatory cortical (GC) response dynamics culminate in decision-related firing

Mahmood, et al. , 2022

Slide 60

Palatability state probability

Ensemble state transitions reflect taste decision-making

Sadacca et al , 2016

Gaping period!

Slide 61

Testing this characterization of GC taste processing

By manipulating palatability

Fontanini & Katz, 2006 (Reduction of arousal)

Grossman et al ., 2008, Moran & Katz, 2014 (taste aversion)

By challenging identity

Yoshida & Katz, 2010 (Taste discrimination [2AFC])

By directly changing gape latency

Li et al., 2016 (cuing aversive taste)

By perturbing GC activity

Mukherjee et al., 2019 (0.5 sec optogenetic silencing)

Slide 62

Sadacca , et al ., 2016

Mahmood, et al. , 2022

Single-neuron responses change suddenly when looked at in this way

Slide 63

Session 1

Session 2

Svedberg, in prep

Scariest fact: taste response dynamics evolve over brief experience

Trial order: top to bottom

Slide 64

This evolution impacts taste response discriminability

Chance

Svedberg, in prep

Slide 65

Sadacca et al , 2016, Baas-Thomas*, Mahmood*  et al. 2026

Ensemble state transitions reflect taste decision-making

Gaping Rat [IMAGE] | EurekAlert! Science News Releases

Palatability state

Gape

Also true for ingestive behaviors

Slide 66

GC is causally involved in initiation of behavior

GC

Mukherjee et al , 2019

None

Before

Transition

After

Transition

![A diagram of a brain

AI-generated content may be incorrect.](./media/ppt/media/image62.png “Google Shape;393;p30”)

Slide 67

Peri-transition time ( ms )

Lin et al, 2021

BLA is necessary for sharp transitions in GC

Onset of palatability response

Intact

BLA-> GCx

Intact

GC

BLA