How a Tiny Cluster of Deep Brain Cells Drives Avoidance

Optogenetic research shows how deep brain structures control anxiety.

Posted Feb 05, 2018

Avoidance is an adaptive response to threat, allowing animals to stay away from dangerous situations, once we learn what they are. While avoidance is often critical for survival, it can backfire as a defensive response when it leads us to avoid situations which are useful and do not post an actual threat. This can be seen clinically in people with Post-Traumatic Stress Disorder (PTSD) and Avoidant Personality Disorder (AvPD), where people fail to engage when it would be helpful to do so.

For example, someone with PTSD might avoid driving or riding in a car after a car crash, perhaps first avoiding the specific model of car in the crash but then avoiding more and more forms of transportation (in extreme cases) due to fear generalization, until they essentially no longer will use any transportation, perhaps even remaining home bound. People with AvPD (similar to Social Anxiety Disorder but more severe and pervasive) have extreme social inhibition, avoid social interactions, feel insecure and inadequate, and are highly sensitive to evaluation by others. Avoidance can also limit what we are able to think about. People often suppress thoughts and inhibit feelings ("experiential avoidance"), which leads to day-to-day problems and interferes with personal growth and development.

When people have maladaptive responses to perceived threat, so much so that it interferes with social relationships and the pursuit of desired activities, avoidance can be crippling. To add insult to injury, avoidance keeps people from engaging in constructive behaviors, such as doing the activity they are afraid to do (e.g. riding in a car, being more socially effective in work and personal settings) when doing so would lead to unlearning the fear response. Avoidance prevents both learning new behaviors as well as being able to return to approaches which worked well—before a negative experience led to an avoidant response. Avoidance can therefore lock-in post traumatic reactions, preventing therapeutic re-engagement and recovery.

The general model in trauma theory, admittedly over-simplified, is to view the amygdala as generating fear (though it is involved with other emotional states) and the hippocampus as providing context, involved with narrative or episodic memory and spatial orientation. So in PTSD, the amygdala would be too active (e.g. all cars would set off the alarm) and the hippocampus would be off-duty, letting us believe that all cars were actually threats (even if we intellectually realize that isn't true), leading to a kind of "brainjack" of higher functions by older, deep parts of the brain.

In contrast, without PTSD someone who'd been in a car crash would be able to recognize they felt afraid to get in a car (if they did), but put it in perspective, recognizing that their fears were exaggerated if understandable. So the basic model has been that in pathological states, there is an imbalance where the amydala is too strong and the hippocampus is too weak, overwhelming the frontal cortex (which is involved with executive functions) and leading to maladaptive avoidance. Therapeutic efforts are directed at restoring that balance, via various means.

However, there is more to the story, as current research by Jimenez and colleagues (2018) elegantly demonstrates. Going beyond the view of the hippocampus as being purely about contextualization of memory and fear, prior work has demonstrated that while the upper side ("dorsal") of the hippocampus is involved in place (context), the belly side ("ventral") is involved with anxiety processing and subsequent behavioral responses. The ventral hippocampus connects (sends "projections" of neurons to) various key brain regions, including the amygdala, hypothalamus (involved with basic physiological activities, stress responses, and fundamental mammalian behaviors), and others.

In order to pinpoint the role of anxiety cells in the ventral hippocampus, Jimenez and colleagues used an optogenetic mouse model. They implanted a tiny microscope deep into the brains of these mice to directly look at cell activity in areas of interested, and used a virus to program those cells so that they could be switched on and off using a small fiber optic cable shining light on that part of the hippocampus (a technique known as "optogenetics"). Because these brain areas are essentially shared by all mammals, and are highly conserved by evolution, findings in this rodent model are likely to apply to human beings in many significant ways.

Jimenez et al., 2018
Looking at anxiety cells in the hippocampus.
Source: Jimenez et al., 2018

This extremely delicate and beautiful approach allowed researchers to see exactly what happened when the mice were exposed to a threatening, stressful situation, and to see what happened when they turned those cells on and off. Would mice showing fear-based responses continue to show those responses when those cells in the ventral hippocampus were switched off? Furthermore, what parts of the brain are activated by the anxiety cells in the hippocampus, leading to avoidance and related reactions to threats?

First, the mice were taught avoidant and fear-based responses using a few standard research protocols (e.g. non-fear based avoidance of a very bright light, fear conditioning using an electric shock box). Researchers compared hippocampal cell responses in the different conditions to make sure they were really seeing responses to anxiety. They found that neurons called vCA1 were selectively activated by fear, leading to avoidance, and not other conditions.

By using optogenetics to switch these vCA1 cells (and different non-anxiety-related cells controlling other responses) on and off, along with a slew of additional manipulations, they were able to determine that not only were these cells specific to anxiety, but also that they controlled fear-related avoidance by sending a message to the lateral hypothalamus, which then generated the behavioral and physiological responses. In other words, when the vCA1 cells were allowed to functional normally, they drove fear-based avoidant reactions and stress responses by activating the hypothalamus.

When they switched off vCA1 cells, the animals did not show avoidant fear reactions, even when they had been conditioned to do so. They also showed that while the same vCA1 cells connected with the amygdala, they did not control avoidant behaviors via this part of the brain. Rather, the connection between the amygdala and hippocampus has more to do with learning fear-based responses in the first place, in keeping with current understanding.

While clinical application in human beings are a long way off, the finding that a specific group of hippocampal cells drives fear-related avoidance via specific influences on the hypothalamus is a fundamental discovery. If we can develop approaches to affect this area specifically, it might be possible to directly target maladaptive avoidance, and other fear-based responses found in clinical disorders via conventional medication approaches, as well as via brain stimulation techniques. In principle, findings like this can also be useful in forensic settings to see whether a reported reaction is "really" happening, by observing what is going on in the brain itself. For example, in theory it is possible to verify on a neurological level a report someone who reports they cannot go to work because of avoidance after an accident.

In addition to clinical potential, having a refined understanding of where vCA1 cells are located, and what they are doing, permits researchers to study human beings more effectively using non-invasive techniques such as neuroimaging studies to find out if the human version of vCA1 cells do the same thing they do in mice. This could be useful diagnostically. For instance, the "biomarker" of higher activity in the ventral hippocampus could be combined with other related findings to lead to a reliable diagnostic test. More and more, especially with psychiatric conditions and other areas where there is no one biological test, using intensive computational methods to make sense of "big data" is the new paradigm.

A good example of this approach is the use pharmacogenomic testing to predict medication response, for example with anti-depressants, anti-psychotics, pain medications, and other treatments. Pharmacogenomic testing is already used in clinical care, and while early on, is becoming a standard. Rather than having one test which tells us whether or not someone will respond to a given treatment, analyzing the results of multiple tests (each of which is of limited use alone) provides clinically meaningful information. The more small tests are available to be built into the computational model, the more useful the overall diagnostic test becomes, and as new research is available, the model can be modified and refined.

Using brain imaging for diagnostic purposes requires being able to make sense of big data as well, is being investigated for other clinical conditions including depression. For example, by looking at imaging data from a group of depressed patients, researchers have been able to identify four different "biotypes" of depression. Next steps would involve developing a clinically-useful diagnostic test, and correlating different biotypes of depression with available treatments to optimize clinical decision-making. As it stand, while the landscape is shifting


Jimenez CJ, Su K, Goldberg R..., Paninski L, Hen R & Kheirek MA. (2018). Anxiety Cells in a Hippocampal-Hypothalamic Circuit. Neuron 97, 1–14 February 7, 2018