Abuzar Mahmood

Aversive taste behaviors, such as gaping, are commonly viewed as fixed, hard-wired responses important for ejecting potentially toxic tastes from the mouth. Yet, taste responses are highly susceptible to modulation by experience and context; for instance, conditioned taste aversion (CTA), a form of learning in which rats are made to respond aversively to a sweet taste after it has been paired with gastric malaise, can cause gaping to previously acceptable tastes. Here, we compare aversive responses to saccharin sodium (sacc CS) caused by repeated aversive experience (in the form of CTA training) to those caused by naturally aversive quinine HCl, and how these responses are aligned with gustatory cortical (GC) activity. Using simultaneous single-neuron ensemble recordings in GC and electromyography (EMG) recordings in the jaw-opener muscle in awake, freely-moving rats and a machine-learning based behavioral classifier, we not only identified two classes of gaping, but characterized how these classes reflect different mechanisms driving aversive behavioral outputs. The first gape class is taste-specific and emerges following GC ensemble dynamics associated with taste identity and palatability coding, leading to canonical gaping responses to naturally or learned-aversive stimuli. The second is rapid-onset (<200 ms following taste delivery), internal state-dependent gaping the animal exhibits to all tastes in a state of heightened vigilance. This generalized aversion is subsequently suppressed once palatability-related responses emerge in GC, such that only gapes to truly aversive tastes remain. Together, these findings demonstrate that gaping is not a unitary reflex but is instead driven by multiple aversion judgment systems active in different contexts.

Upon delivery of a taste onto the tongue, gustatory neural activity determines whether the stimulus is ingested or rejected. While some work in rodents has been devoted to investigating the neural activity leading to the rejection decision and its associated orofacial movements, little is known about what behaviors lead to ingestion of palatable tastes (and what neural activity is associated with that decision), largely because identifying ingestion-related behaviors is a difficult challenge—and probably undoable with video analysis given that the behaviors are largely intraoral. To address this gap in our understanding, we analyzed simultaneously-collected electromyographic (EMG) activity of the jaw opener muscle and the firing of gustatory cortical (GC) ensembles. We developed a machine-learning classifier to identify individual orofacial movements from EMG signals, demonstrating that it outperforms previously developed methods and using the technique to reveal three novel subtypes of ingestion-related tongue/mouth movements. Investigating the dynamics of these behaviors, we found that the frequency of occurrence of each type subtype shifts significantly at the time of the consumption decision, and is both tightly coupled with and reliably follows the transition in GC population activity into the state reflecting the tastant’s emotional/hedonic value. However, rather than the onset of single “ingestion” movement (as occurs for rejection decisions in the form of gapes), we show that the transition to ingestion is instead characterized by a collective change in the frequencies all ingestion-related behaviors. These findings demonstrate a direct link between neural dynamics in GC and the orchestration of the physical movements that define ingestive behavior, highlighting GC’s general role in taste perception, decision making, and the control of motor actions.

Conventionally, taste circuitry is considered feedforward, traveling up from the brainstem, with each additional node containing more sophisticated information. We challenge this convention by demonstrating that amygdala and cortex instead influence each other bidirectionally, yet in a direction-specific and asymmetric manner. These influences appear to drive different parts of the taste response, with distinct patterns for decision-making and behavioral output; furthermore, involvement in cortex-amygdala functional connectivity determines the strength of encoding in cortical single neurons.

Upon delivery of a taste onto the tongue, gustatory neural activity determines whether the stimulus is ingested or rejected. While some work in rodents has been devoted to investigating the neural activity leading to the rejection decision and its associated orofacial movements, little is known about what behaviors lead to ingestion of palatable tastes (and what neural activity is associated with that decision), largely because identifying ingestion-related behaviors is a difficult challenge—and probably undoable with video analysis given that the behaviors are largely intraoral. To address this gap in our understanding, we analyzed simultaneously-collected electromyographic (EMG) activity of the jaw opener muscle and the firing of gustatory cortical (GC) ensembles. We developed a machine-learning classifier to identify individual orofacial movements from EMG signals, demonstrating that it outperforms previously developed methods and using the technique to reveal three novel subtypes of ingestion-related tongue/mouth movements. Investigating the dynamics of these behaviors, we found that the frequency of occurrence of each type subtype shifts significantly at the time of the consumption decision, and is both tightly coupled with and reliably follows the transition in GC population activity into the state reflecting the tastant’s emotional/hedonic value. However, rather than the onset of single “ingestion” movement (as occurs for rejection decisions in the form of gapes), we show that the transition to ingestion is instead characterized by a collective change in the frequencies all ingestion-related behaviors. These findings demonstrate a direct link between neural dynamics in GC and the orchestration of the physical movements that define ingestive behavior, highlighting GC’s general role in taste perception, decision making, and the control of motor actions.

Gustatory cortical (GC) and basolateral amygdalar (BLA) taste responses consist of an inter-regionally coherent 3-part state sequence. This coherence suggests that reciprocal BLA-GC connectivity is important for taste processing, but it remains unknown: 1) whether BLA-GC coherence actually reflects a reciprocal "conversation" (as opposed to one region simply driving the other); and 2) whether such a "conversation" has anything to do with the taste processing observed within GC response dynamics. Here, we address these questions using network and single-neuron analysis of simultaneously-recorded GC and BLA taste responses in awake rats. We find asymmetric, reciprocal μ-frequency influences that reflect taste processing dynamics: BLA→GC influence dominates between 300 and 1000msec (the epoch in which BLA codes palatability); afterward, when GC responses become palatability-related and GC has been shown to release a behavior-relevant signal, the direction of influence reverses, becoming GC→BLA. Follow-up analyses demonstrate that this "turn-taking" exists alongside effectively synchronous amygdala-cortical coupling-the two regions functioning as a unified structure. Finally, to assess the implications of these interactions for single-neuron responses, we tested the response properties of GC neurons categorized by their inferred connectivity with BLA: GC neurons influenced by BLA produce stronger taste-specific and palatability-related responses than other GC neurons, and the strongest taste encoding is specifically found in GC neurons that both influence and receive influence from BLA-those most deeply embedded in the reciprocal circuit. These results, consistent with findings in multiple systems, support the novel conclusion that taste processing and decision-making is a function of the amygdala-cortical loop.