Avoiding Loss: Neural Key To Anxiety And PTSD

People are often willing to take risks to make a profit, but when there's a chance of losing, they will do anything they can to avoid the same risks - even if it means acting irrationally. This decision-making bias is at the heart of prospect theory, for which Daniel Kahneman received the 2002 Nobel Prize in Economic Sciences. Yet how the risk of loss biases learning and decision-making - and the neural basis for this bias - remain largely unknown.

A new study - whose findings have been published in two papers, in Nature and Current Biology - uncovers brain mechanisms that make us more sensitive to losses and cause us to award them greater weight than gains during learning and decision-making. Excessive or inappropriate activation of these mechanisms may help explain behaviors seen in people with anxiety disorders and PTSD.

The study was led by Prof. Rony Paz and Dr. Tamar Reitich-Stolero of the Weizmann Institute of Science's Brain Sciences Department and neurosurgeon Prof. Ido Strauss and neurologist Dr. Firas Fahoum from Tel Aviv Sourasky Medical Center (Ichilov).

How the brain weighs risk

The amygdala - an almond-shaped structure deep in the brain's temporal lobe - is key to processing fear and stress and regulating our response to threats. "In recent years, physicians have been increasingly implanting electrodes into the amygdala and other deep brain regions of patients with severe, drug-resistant epilepsy in an effort to locate the origin of their seizures," Strauss says. "Unlike EEG electrodes placed on the scalp, which record the average electrical activity of millions of neurons in large portions of the brain, depth electrodes can detect the firing of individual cells. This allows researchers to study how small groups of neurons compute information in real time."

When facing potential losses, people are less likely to rely on accumulated knowledge and more likely to keep exploring alternatives

In the new study, Weizmann Institute researchers joined forces with Ichilov physicians and used depth electrodes to probe the neural mechanisms of learning and memory under the risk of loss. In the first part of the study, published in Nature , participants dealt with two types of trials while activity of single neurons in their brains was being recorded: Some trials offered the chance to gain points, others involved the risk of losing points. A sound alerted participants to the type of trial - gain or loss, followed by a display of two geometrical shapes - one indicating that the chances of gain (in the gain trial) or loss (in the loss trial) were high, the other, that the chances were low. Over time, participants learned which options consistently led to better outcomes.

"Even so, performance was different in the two types of trials," Reitich-Stolero says. "In the loss tasks, participants sometimes ignored the optimal choice and kept desperately searching for strategies that might prevent losses altogether, not just minimize the risk of loss. In contrast, in the gain tasks they stuck with the best option once they had learned it and were less likely to look for creative solutions."

Every living creature is caught in a constant struggle between using existing experience and searching for new and better ways. The new study shows that humans deal with this dilemma differently in gain and loss situations. When facing potential losses, people are less likely to rely on accumulated knowledge and more likely to keep exploring alternatives - even if those choices bring more losses and less gains in the short term.

To understand the neural mechanisms behind this behavior, the researchers monitored the activity of hundreds of single neurons across different brain regions. They identified a subgroup of cells in the amygdala and temporal cortex that fired at higher rates just before participants decided to try a new strategy. These signals essentially indicated whether the person was about to explore new options or rely on past knowledge.

"This mechanism was equally active for loss and gain tasks, so at first, we couldn't understand - why do people explore more in loss situations?" Reitich-Stolero says.

The researchers hypothesized that neural noise - random fluctuations in the neurons' rate of firing - might play a role. Indeed, they found higher levels of neural noise in the amygdala when participants faced potential losses and when they searched for new strategies to avoid loss. Computational modeling showed that this noise was linked to feelings of uncertainty, and that people were more sensitive to uncertainty when faced with the prospect of losses than with that of gains, which, in turn, made them more likely to seek new strategies. "When exploratory behavior becomes uncontrolled, people can get stuck in a constant search for better options - a hallmark of anxiety disorders," Reitich-Stolero says.

When the brain overgeneralizes

The second part of the study, published in Current Biology , examined another aspect of the decision-making process: the ability to generalize. Participants heard tones they had previously learned to associate with gains or losses, as well as new tones that were similar or different. The researchers found that participants tended to overgeneralize in the loss situation, treating a broader range of new tones as "familiar" when they were similar to a tone previously associated with losses. Moreover, their brains interpreted these "loss tones" as if they signaled danger.

"The ability to generalize is an essential element of intelligence," Paz explains. "It developed over the course of evolution, allowing us to create broad safety rules based on past experience and protecting us from new threats. But when generalization gets out of control, it can be harmful. That's why, after we hear a rocket siren, even the sound of a passing motorcycle can trigger panic. It's a great defense mechanism, but when it's overactive, as in PTSD, it can lead to stress, anxiety and depression."

Recordings from intracranial electrodes revealed a neural mechanism that explains why we are more likely to generalize when there is a risk of loss. Neurons in the amygdala were overactivated by tones similar to the "loss tone," and this activity preceded the generalization response. In other words, the amygdala's heightened activity biased the brain to interpret new sounds as threats.

"Based on the level of neural activity, we could predict whether a person would make a generalization and misidentify a tone as familiar," Paz notes. "Negative conditioning can make us believe that we heard a different tone from the one that was actually sounded - that is, it can alter our sensory perception. While the amygdala's role in fear and anxiety has been known for years, only recently have we been able to study decision-making in humans at such high resolution. Now we can better understand what goes wrong in various disorders - and point the way to improved treatments for post-trauma and mood disorders."

Study authors also included Dr. Kristoffer C. Aberg, Dean Halperin and Carmel Ariel from Weizmann's Brain Sciences Department; Dr. Genela Morris from Tel Aviv University; and Drs. Lilach Goldstein and Lottem Bergman from Ichilov.