X hits on this document

17 views

0 shares

0 downloads

0 comments

3 / 8

a

Conditioned fear

Cortically distributed (AI, ACC, hipp.)

representation of the CS-US/pairing

cs us

CE

Visual cortex

B

LA cs us

Primary (SI) and secondary (SII) somatosensory cortex

Visual thalamus

Somatosensory thalamus

cs

Autonomic output

us

b Observational fear

Cortically distributed (AI, ACC, hipp.)

representation of the CS-US/pairing

cs

MPFC: mentalizing

CE

ACC, AI Empathetic emotion

Visual cortex

B

LA cs

Visual cortex

Visual thalamus

Visual thalamus

cs

Autonomic output

Social US

c

Instructed fear

Left hemisphere: language representation Threat

Left hemisphere: cortically distributed (AI, ACC, hipp.) representation

of the CS-threat/pairing

cs Threat

CE

Visual cortex

B

LA cs

Visual thalamus

cs

Autonomic output

Jessica Iannuzzi

Given the adaptive function of the ability, it is not surprising that many animals, including birds33, mice34, cats35, cows36 and primates37–43, can learn fears by observing a conspecific. In one ecologically valid study34, model mice were attacked by biting flies while observer mice watched. When exposed 24 hours later to flies, whose biting parts had been removed, the model and observer mice expressed conditioned analgesia and avoidance responses to similar degrees, implying that individual and social fear learning were equally effective. The strength of the model’s fear response during its individual learning was not correlated with expressed fear learning in the observer at a later test. In contrast, such a relationship was found during observational fear learning in pri- mates39, indicating that there may be a greater reliance on emotional expressions during the learning process. The rich and flexible muscu- lature of the primate face allows it to produce a wide repertoire of emotional expressions, superior to that of many other species44. The cortical areas dedicated to face processing are also relatively enlarged in primates45, implying an enhanced ability to rely on facially transmitted

emotional information. In monkeys39,40 and humans

41–43,46, facial fear expression is a

reliable US. Cage-reared monkeys were shown either live presenta- tions or movies of model monkeys reacting fearfully to snakes (toy or real) or to non–fear-relevant objects39. When fear-relevant objects were used, the relationships between the strength of a learning

NATURE NEUROSCIENCE VOLUME 10

[

NUMBER 9

[

SEPTEMBER 2007

REVIEW

Figure 2 A neural model of nonsocial and social fear learning in humans. The arrows describe the flow of information between different functional brain regions. Although the arrows point only in one direction, the connectivity might be bidirectional. (a) Fear conditioning occurs by associating the visual representation of the CS with the somatosensory representation of the aversive US. The lateral nucleus (LA), in which sensory representations of the CS and US converge, is believed to be the site of learning. The amygdala also receives input from the hippocampal memory system (hipp.), anterior insula (AI) and anterior cingulate cortex (ACC) containing secondary representations of the CS and US, information about the learning context and the internal state of the organism. (b) In observational fear learning, the visual representation of the CS is modified by its association with a representation of the distressed other, serving as the US. As in fear conditioning, it is hypothesized that representations of the CS and the US converge in the LA. The strength of the US may be modified by MPFC input related to the interpretation of the other’s mental state, as well as cortical representations of empathic pain through the ACC and AI. (c) Instructed fear learning occurs by modifying the processing of the visual representation of the CS through its association with an abstract representation of threat. Instead of being coded in the amygdala, the CS–‘threat’ US contingency is likely to be represented in a cortically distributed network, critically depending on the hippocampal memory system.

model’s expressed distress, the observer’s immediate response to the model’s distress, and the resulting fear learning in the observer were comparable to the relationship reported between US, UR, and condi- tioned response in classical fear conditioning39,40. This one-trial social encounter with a fearful model produces a robust fear response that lasts several months39. Again, these findings strongly indicate that observational fear learning draws on the same processes as fear conditioning. Still, the neural processes remain to be explored in nonhuman animals.

The ultrasocial environment of humans provides ample opportu- nities to watch others’ emotional responses to stimuli47,48. Children with subclinical animal phobias or extreme fears toward certain situations, such as darkness, often report having observed parents fearful in the same or similar situations49,50. Normal children can acquire a strong and persistent aversive response to a fear-relevant object (such as a toy snake) after seeing it paired with their mothers’ fear expressions46.

In adults, another person’s arm movement in response to a shock can act as an US, but only when the observer believes that it was caused by a shock, not when the model’s arm moves without a shock or when a shock is delivered without arm movements37. These results support the conclusion that perceptual properties of the learning model interact with the observer’s knowledge to instigate an unconditioned response. Similarly, information about another person’s spider phobia can induce an aversive response to a spider that is presented to the allegedly phobic model, even without any physical cues of distress38, and the affective response in an observer can be modified by context51,52. In sum, research on observational fear learning consistently establishes that a facial expression can serve as an US, but, as discussed below, social variables can also modulate the response.

Observational fear learning may draw on the same processes as fear conditioning, with the expression of the conspecific learning model serving as the US. However, some studies on social learning in rats do not replicate core features of classical fear conditioning, such as blocking, overshadowing and latent inhibition53, and at least one study did not find evidence for fears acquired through observation54. In contrast, humans demonstrate classical conditioning characteristics for observational learning, including overshadowing and blocking55, indicating that observational learning may show greater interspecies variability than does classical fear conditioning.

1097

Document info
Document views17
Page views17
Page last viewedSun Dec 04 10:09:35 UTC 2016
Pages8
Paragraphs417
Words8422

Comments