Attending to the Amygdala’s Hold Over Aversive Salience
Material below summarizes the article Basolateral Amygdala Neurons Maintain Aversive Emotional Salience, published on March 21, 2018, in JNeurosci and authored by Auntora Sengupta, Joanna O.Y. Yau, Philip Jean-Richard Dit Bressel, Yu Liu, Zayra E. Millan, John M. Power, and Gavan P. McNally.
The ability to learn about and respond to sources of danger is essential to survival. A variety of lines of evidence, ranging from single-unit recording studies in rodents to functional neuroimaging or neuropsychological studies in humans, show the amygdala is critical for this learning.
Fear learning can be studied in laboratory animals using Pavlovian fear conditioning. The experimenter arranges a conditioned stimulus (CS) (e.g., an auditory stimulus) to signal delivery of an aversive event (e.g., shock to the paws).
The consequence of these pairings is the animal will show fear responses to the CS when it’s subsequently encountered. This fear learning is readily acquired, often within a few trials, and it persists for a long time.
Despite this procedural simplicity, fear learning is not a simple process. It involves complex psychological processes of attention, stimulus selection, error-detection, and stimulus processing.
We know from past work that principal neurons in the basolateral amygdala are sensitive to both predictors of danger (conditioned stimulus) and the danger itself (unconditioned stimulus), and that these neurons undergo complex increases and decreases in activity during fear learning.
However, we know little about how the activity of amygdala principal neurons at specific times of fear learning relate to the specific psychological processes that mediate learning.
In our experiments, we combined fiber photometry, optogenetics, and sophisticated behavioral approaches in an attempt to solve this problem.
Neurons transfected with the genetically encoded calcium sensor gCaMP6 show calcium-mediated fluorescence as an indicator of cellular activity. The advantage here is distinct neuronal populations can be targeted depending on the virus promoter and/or rodent line. Fiber photometry records this bulk calcium-dependent fluorescence via an implanted optic fiber.
In this way, we studied the activity of basolateral amygdala principal neurons during pairings of a conditioned stimulus and unconditioned stimulus (US). Consistent with the literature, presentations of the shock US evoked activity in these neurons.
Next, we used optogenetics, a technique that uses light to excite or inhibit genetically modified neurons, to silence amygdala principal neurons during presentations of the shock US, the period in which they showed the greatest increases in activity. Again, consistent with past work, we found impairments of fear learning.
These two findings, although interesting and the same as those typically found in the literature, do not isolate the learning functions served by US-related activity of principal neurons. So, our subsequent experiments used behavioral approaches to isolate the function of these neurons.
First, we inhibited principal neurons during fear extinction, specifically during moments of non-reinforcement (that is, the time when the shock was expected to occur but didn’t).
Optogenetic inhibition facilitated loss of fear during extinction training. Moreover, it had a long-lasting impact because it protected this extinguished fear from later showing relapse.
This pattern of findings — impaired fear learning and augmented fear extinction — is broadly consistent with a class of theories arguing the primary function of US-related activity in the amygdala during fear learning is to encode a prediction error signal that instructs association formation.
This error signal reports the difference between the US that actually occurs and the US that is predicted based on the conditioned stimuli present on the trial. However, this pattern of findings is not uniquely explained by these theories. Other explanations are possible.
A prediction error account states optogenetic inhibition should also facilitate other forms of inhibitory learning. So, we studied the effects of optogenetic inhibition in a behavioral design called conditioned inhibition or safety learning.
We trained rats that a CS signalled the absence of shock, transforming this CS into a safety signal able to inhibit fear responses. Again, we silenced principal neurons only during the moments of non-reinforcement necessary for safety learning. Optogenetic inhibition impaired, not facilitated, safety learning.
Our results show the activity of basolateral amygdala principal neurons during reinforcement and non-reinforcement does not simply reflect a prediction error signal. If it had, optogenetic inhibition should have augmented safety learning like it augmented extinction learning.
This was not the case. Instead, our results suggest an even more fundamental role for the activity of principal neurons during fear learning.
This role can be understood by imagining a simple behavioral experiment: habituation. We repeatedly present an auditory stimulus to an animal. Initially, the animal orients towards the source of the sound, but this responding declines across repeated presentations of the stimulus. The animal comes to ignore the stimulus and not respond to it. The stimulus has lost its salience — its ability to command attention and control behavior — because it signals nothing of relevance or importance.
Such habituation is clearly adaptive. It reduces demands on our attentional and learning systems and allows us to safely ignore innocuous events in the world. However, such habituation would be maladaptive if, for example, the auditory stimulus was followed by a dangerous or painful event. Now, the stimulus should command our attention and control our behavior.
Our findings suggest a primary function of amygdala principal neurons during fear learning, specifically during moments of reinforcement and non-reinforcement, is precisely this maintenance of CS salience, so that it’s able to command attention and control behavior.
Without this activity, the CS becomes divorced from the fear system. In our experiments, optogenetic inhibition caused a reduction in salience of the CS so that it was more likely to be ignored, harder to associate with fear or safety, and less likely to control behavior.
Salience maintenance is a necessary precursor for formation of an association between the CS and US. However, it’s not association formation itself.
So, our work suggests a more fundamental role for the amygdala in fear learning than previously thought. It’s instructive to consider whether the effects of other amygdala manipulations can also be explained as influences on CS salience.
Moreover, our findings are relevant to human anxiety. One of the goals of treatments for human clinical anxiety is to reduce the ability of threat-related stimuli to command attention and control behavior. Our findings suggest approaches that reduce the salience of such stimuli may be especially effective.
Basolateral Amygdala Neurons Maintain Aversive Emotional Salience.Auntora Sengupta, Joanna O.Y. Yau, Phillip Jean-Richard Dit Bressel, Yu Liu, Zayra E. Millan, John M. Power, and Gavan P. McNally. JNeurosci Oct 2017, 2460-17; DOI: 10.1523/JNEUROSCI.2460-17.2017.