Nature Neuroscience pp1500 - 1502

A major risk factor for schizophrenia is a genetic mutation called 22q11.2 microdeletion, which occurs in 1 of 4000 people. One third of people with this mutation develop schizophrenia or another psychotic disorder. Two groups shed light on the linkage between the mutated genes and the cognitive and physiological changes that may be involved in schizophrenia, in the November issue of Nature Neuroscience.

In one study, Doron Gothelf and colleagues administered psychological tests to children with and without 22q11.2 syndrome and examined them again several years later, in late adolescence or early adulthood. They found a strong effect of a particular mutation in the gene COMT, which produces a protein that breaks down the neurotransmitter dopamine. Subjects with this mutation had an abnormal decrease in the size of the prefrontal cortex (a brain region involved in reasoning and control), as well as lower IQs and more frequent psychotic symptoms than subjects without the mutation.

In the other study, Maria Karayiorgou and colleagues examined a mouse model of schizophrenia that contains a mutation in the gene for the enzyme proline dehydrogenase (PRODH), which breaks down the proposed neuromodulator proline. They report that this deficiency in PRODH alters the expression of the COMT gene. Interaction between these two genes modulated several schizophrenia-related phenotypes in these mice, such that COMT inhibition exaggerated or induced behavioral deficits in the PRODH-deficient mice. This synergistic interaction between these genes could explain the high disease risk associated with deletions in the 22q11.2 locus.

Nature Neuroscience pp 1603 - 1610

Some patients who suffer a right-hemisphere stroke develop a syndrome called 'neglect', in which they ignore the entire left side of their bodies and of the objects around them. This deficit is caused by abnormal activation in intact areas of the brain connected to the damaged areas, rather than by the original damage itself, reports a paper in the November issue of Nature Neuroscience.

Maurizio Corbetta and colleagues scanned neglect patients using functional magnetic resonance imaging (fMRI) immediately after the stroke, when they were impaired at detecting targets on their left side, and again many months later, when they were much better at the task. The authors found that an undamaged brain area in the right hemisphere did not activate at all in the first scan, but activated strongly later on, when the patients� performance was better. This area, the dorsal parietal cortex, is normally involved in shifting attention, and is connected to the temporo-parietal junction and the prefrontal cortex, areas which are damaged in spatial neglect. There results show that behavioral deficits might result not from actual damage to a brain area, but from alterations in activity in brain areas connected to the damaged region.

Nature Neuroscience pp 1611 - 1618

The perceived moral character of a player's opponent in a trust game influences activity in reward-related brain areas, reports a study in the November issue of Nature Neuroscience. The results also suggest that people rely less on feedback about the outcome of their choices when they have prior information about an opponent's character.

Elizabeth Phelps and colleagues used functional magnetic resonance imaging (fMRI) to track people�s brain activity as they played a game. On each round, participants chose whether to keep a monetary reward or to share it with an opponent. If their opponent also chose to share the reward, participants would get more money than if they had chosen to keep the reward initially. But if their opponent kept the reward when they chose to share, participants would receive no money. Sharing a reward was therefore a risky choice, which could result in a greater reward or a complete loss.

Before the game, participants read descriptions of their opponents� life events that portrayed the opponent in a positive, negative or neutral context. The opponents� responses during the game were the same for all three types of opponent. However, subjects made the risky choice to share their reward with the 'good' opponent significantly more often. Activity in the caudate nucleus - part of the brain that normally differentiates between good and bad feedback - signaled feedback information only for the 'neutral' partner. These findings suggest that 'trial and error' learning can be modified by social context.