Vikaas Sohal and his colleagues set out to test the hypothesis that excessive entry of calcium into neurons within the medial prefrontal cortex (mPFC) impairs social behavior and contributes to additional aspects of autism. The researchers learned that the excessive entry of calcium does not affect all neurons in the mPFC equally. Rather, neurons in the mPFC can be divided into different subpopulations, and the excessive entry of calcium has a profound effect on one of these subpopulations.

This observation led Sohal and his group to wonder whether autism affects all neurons in the mPFC equally, or whether distinct changes occur in different subpopulations of neurons within the mPFC.

A closely related question is whether all neurons in the mPFC contribute to abnormal behavior in autism or whether there are certain subpopulations of mPFC neurons that play a particularly important role in autism-related behaviors. Answering these questions could help the researchers to focus on specific classes of neurons and detect changes that might be missed by lumping all mPFC neurons together.

Sohal and his team studied three distinct mouse models of autism. In every case, the researchers found that the properties of mPFC neurons do not change in a uniform way. Rather, different subpopulations of mPFC neurons exhibit distinct changes.

The researchers also found one specific subpopulation of mPFC neurons that is altered in a consistent way across all three mouse models of autism, in each case making those neurons less responsive to input.

Sohal and his colleagues also carried out experiments in which they stimulated specific subpopulations of mPFC neurons in normal mice. The researchers disrupted activity in the same subpopulation of mPFC neurons that exhibit decreased responsiveness in mouse models of autism. Remarkably, this leads to behavioral deficits resembling aspects of autism. Specifically, interfering with the normal activity of these neurons causes mice to engage in less social interaction and to exhibit less flexibility while learning a task.

By contrast, altering activity in other populations of neurons does not produce the same effects. These results suggest that autism may involve distinct changes in different subpopulations of neurons within the mPFC and that certain subpopulations may play a particularly important role in autism.