Welcome to

Mind Matters

Sciam.com's "seminar blog" on the sciences of mind and bbrain. Each week, top researchers in neuroscience, psychology and psychiatry explain and discuss the research driving their fields. Readers can join them. We hope you will.

This week we ponder

Can We Control Our Fears?

For years, the amygdala (shown above in red) has been considered "necessary and sufficient" for the learning and expression of conditioned fear. A new study suggests it has a crucial anatomical ally.



by David Dobbs, Editor, Mind Matters

For some two decades, ever since Joseph LeDoux and others began elucidating the neural circuits by which we learn fear, a pair of almond-sized brain bits called the amygdala has been considered pretty much solely responsible for the learning and expression of conditioned fears. This certainly matches the general feeling that fear is a fundamental business resistant to rational control. But is it oversimplified? Or course, no one ever claimed the amygdala stood as an island apart; in the brain as in life, it's all connected. Yet while researchers had illuminated the basic fear circuit quite starkly, demonstrating its workings as plainly as you can a light switch, no one had been able to demonstrate that sort of binary, togglelike function in any connections between the amygdala and other brain areas. This ended with the publication of the paper reviewed below. As its title suggests, "Activity in Prelimbic Cortex is Necessary for the Expression of Learned, But Not Innate, Fears" (by Kevin Corcoran and Gregory Quirk of the Ponce School of Medicine in Ponce, Puerto Rico, from the Journal of Neuroscience, 24 Jan 2007) shows that at least in the rat, the prefrontal cortex -- the brain region most responsible for managing our behavior -- plays some crucial and surprising roles in learning and expressing conditioned fear.


Can We Control Our Fears?

by Sevil Duvarci and Denis Paré Center for Molecular & Behavioral Neuroscience Rutgers University, Newark, N.J..

Every organism's survival depends on its ability to respond to danger. The solution retained by natural selection for this purpose is fear, an integrated pattern of responses (behavioral, autonomic, and so on) generated reflexively in the face of threatening stimuli. Neuroscientific work in the last decade has revealed that the neural circuitry underlying fear is highly conserved in mammalian species, from rats to humans; that is, the fear mechanisms and systems developed by the earliest mammals have been carried over through evolution to those living today. Within these systems the amygdala, a pair of almond-shaped structures deep in the temporal lobe, plays an essential role. It is central both in triggering innate fear responses, such as the fear of predators or loud, unexpected noises, and helping to build new fear responses through experience, as when we learn the hard way not to touch a hot stove. Until recently it was thought that fear learning depended on changes taking place in the amygdala, independently of inputs from the cerebral cortex. In other words, our elemental fear circuits, whether the fear was innate or learned, did not need or use much input from the cortex. Now a study published recently in the Journal of Neuroscience by Kevin A. Corcoran and Gregory J. Quirk challenges this view, disclosing a previously unsuspected difference between circuits responsible for learned versus innate fear. Learning to Fear The approach commonly used to study fear learning in the laboratory is classical fear conditioning: a rat learns to fear a neutral sensory stimulus, such as a tone, after the neutral stimulus has been paired to a noxious stimulus, generally a mild electrical shock to the feet. The fear response typically seen and monitored in such studies is behavioral freezing: the arrest of all movements other than breathing. In the wild, freezing enhances the likelihood that a rat will survive an encounter with a cat because this predator is highly sensitive to moving stimuli. Rats also learn to freeze in response to other stimuli that become associated with danger. Fear learning thus enables animals to predict and avoid dangerous situations on the basis of prior experience. To see whether cortical inputs affect fear responses, Corcoran and Quirk locally injected a small quantity of tetrodotoxin (TTX), a fish toxin that temporarily blocks neural activity, in the prelimbic cortex, a cortical area that is reciprocally connected with the amygdala. Earlier work had revealed that amygdala and prelimbic neurons show increased responses to conditioned stimuli as a result of conditioning. Thus it seemed plausible that chemically inactivating the prelimbic cortex with TTX might interfere with fear responses. To test this idea, Corcoran and Quirk used TTX to chemically inactivate the prelimbic cortex and investigated how that affected learned and innate fears. In the learned-fear scenario, they found that if they trained a rat to fear a tone while it was drug-free and then inactivated the prelimbic cortex with TTX just before testing 24 hours later, the animals did not show the normal freezing response to the tone. They froze when they learned to associate the tone with the shock, but not when exposed to it the next day, when normally the tone would make them freeze. Surprisingly, however, when the researchers injected the TTX into the prelimbic cortex just before conditioning and then assessed freezing when the rat was drug-free during memory recall one day later, the rats exhibited normal fear levels and freezing. This response occurred even though the freezing reaction had been much reduced during the training session. The rats had apparently learned during the conditioning session even though they did not show conditioned fear at the time. This finding suggests an interesting dynamic: prelimbic inputs to the amygdala are not needed for acquiring conditioned fear but are required for its normal expression. But what about innate fear? Would the prelimbic cortex be necessary for expressing it too? To explore this question, Corcoran and Quirk tested the effect of prelimbic TTX injections on two types of innate fear responses: a rat's fear of cats, and a rat's fear of open spaces. It is well known that when a laboratory-reared rat sees a cat close at hand for the first time, it will freeze -- even though neither it nor its ancestors have ever seen a cat. The fear is innate. But would the fear be expressed if the prelimbic cortex was inactivated? Corcoran and Quirk chemically inactivated rats' prelimbic cortex with TTX, then placed a cat in an arena that was adjacent to the rats' area but separated by a wire screen. The rats froze, as normal -- a surprise, given the results of the learned-fear experiments, in which inactivation of the prelimbic region prevented fear expression. Corcoran and Quirk got similar results with an "open field test." The open field test exploits the well-established fact that rats fear open spaces: a rat placed in a large enclosure will quickly head toward a wall and then spend most of the time along the walls. Corcoran and Quirk found that rats that received TTX injections in the prelimbic cortex followed this pattern. They did not need their prelimbic cortex to express innate fear responses. Can Learned Fears Be Altered? The big news in this study is that at least some cortical inputs to the amygdala -- those from the prelimbic cortex -- are involved in the expression of conditioned fear. This involvement gives learned fear a previously unrecognized anatomical component. And it establishes that there is at least one difference between the networks underlying the expression of innate and learned fears. These observations have far-reaching implications. First, they suggest that the expression of learned fear is flexible and subject to modulation by the prelimbic cortex, depending on the circumstances; our expression of learned fears is less rigid and less automatic than the expression of innate fears, which are beyond the reach of the cortex. These observations also raise the possibility that hyperactivity in the prelimbic region might contribute to human anxiety disorders that are caused by over-expression of learned fear, such as post-traumatic stress disorder. If that proves true, reducing the activity of the prelimbic cortex might constitute a useful strategy for the treatment of these debilitating disorders, while leaving innate fear responses intact. If learned fear is necessary, so is our ability to control it. This study reveals some dynamics that might be crucial in exercising that control.
Sevil Duvarci and Denis Paré research mechanisms of fear, learning, and memory at the Paré Lab at Rutgers University's Center for Molecular and Behavioral Neuroscience, where Devarci is a graduate student and Paré is a professor of neuroscience.