Model Predicts Neural Inhibition’s Effects on Anxiety

by Kayt Sukel

December 20, 2010

Imagine the smallest decision –what part of the newspaper to read first, what time to leave work to avoid rush hour, what kind of cereal to buy at the market – had the power to stop you in your tracks. People with anxiety often have trouble with these types of seemingly inconsequential decisions. Research reported Sept. 21 in the Proceedings of the National Academy of Sciences suggests this difficulty is due to a glitch in a process called neural inhibition.

Modeling predictions

Hannah Snyder, a graduate student in Yuko Munakata’s lab at the University of Colorado at Boulder, did not set out to study anxiety– she was interested in looking at how the brain selects between similar yet competing options when making decisions. But she and Munakata soon learned this kind of competitive lateral inhibition, or the ability for one neuron to suppress the activity of another when it’s only slightly more active than the other, plays an important role in anxiety.

“One of the hallmark cognitive problems in generalized anxiety disorder is choice overload,” says Snyder. “A person with anxiety will go into a store to buy a birthday card, look at every single card in the store and then leave an hour later without having bought anything. It may seem trivial but it creates a lot of anxiety and has the potential to affect much more serious life decisions.”

Snyder and Munakata, with the help of Randall O’Reilly and other collaborators, built a neural network simulation to test the idea that neural inhibition plays a critical role in decision-making. When neural inhibition was increased, the computation-based model was better at choosing a single option among all possibilities. Conversely, when neural inhibition was decreased, the model fared worse in selecting a single option.

“In the model, we saw neural inhibition allows competition to be resolved between multiple possible responses,” says Snyder. “The selection was not top-down or directed but rather it used a competitive lateral inhibition process that would select an option that was only slightly more active than the others.”

Converging human evidence

The group tested the predictions made by the neural network simulation by running a neuroimaging study. Eighty-five people who did not have a diagnosed anxiety disorder but could be rated as having high or low anxiety levels were asked to select the first verb that came to mind when presented with a noun. The researchers focused on activation levels in the left ventrolateral prefrontal cortex, a brain area thought to be involved in both word choice and neural inhibition.

“Individuals with higher anxiety levels were not only worse at making these decisions but the activity in the left ventrolateral prefrontal cortex was less typical than what we saw in the low anxiety participants,” says Munakata. “That activity looks like some kind of competition between the neurons when we try to choose our words.”

A third study used the drug midazolam, a GABA-agonist that increases neural inhibition, to test whether it would improve selection in both the high anxiety and low anxiety groups. It did just that.

“The model was able to make specific predictions that did bear out in our empirical work,” says Snyder. “If we increased inhibition, we improved selection. By decreasing inhibition, we impaired selection. The studies show that neural inhibition is important in the ability to pull a word to mind when nothing comes easily to mind.”

Implications for treatment

Munakata, Snyder, and O’Reilly caution that this work is preliminary – but it’s a good first step to what may be a new realm of possibilities for treating anxiety, be they drugs or targeted cognitive therapy.

“There are a lot of potential advantages to using a computational mechanistic approach with a disorder like anxiety,” says O’Reilly. “It gives you a lot of precision and more insight into how the neural mechanisms of anxiety may operate. That precision can help with treatment design later.”

These studies looked at a typical population; the group now is looking to see how the findings may apply to peoplewho have anxiety disorders.

“You would think that the worse your neural inhibition, the worse you’ll struggle with decisions – and that should be even more evident in those with extreme anxiety,” says Munakata.

In the meantime, Snyder says that while drugs or therapies may still be decades away, awareness isn’t. “Since we see choice overload plays such an important role, you may be able to help your anxiety by limiting your choices,” she says. “If you can’t change the way your brain works, change your environment.”