New in Neuroscience: Pay Attention to this, not that

listen - ignore switch.jpg

"New in Neuroscience" is a new feature highlighting recent neuroscience findings. In this edition, Malcolm Campbell discusses a recent research article about how the brain pays attention to some things, while ignoring others. 

The ability to selectively attend to relevant stimuli and ignore distractions is essential to animal survival. This ability is especially interesting to neuroscientists since it involves the interaction of executive control--the “mind” of the animal--with early sensory processing. It is widely believed that executive control originates in the cortex, whereas sensory signals originate in the periphery and travel to the cortex through subcortical structures such as the brain stem and the thalamus.

The thalamus is a small potato-shaped structure at the core of the brain where sensory information converges before entering the cortex. Because of this role, it is sometimes called the gateway to the cortex. In this paper, Ahrens et al explore the neural circuitry underlying the gatekeeping of sensory inputs by the thalamic reticular nucleus (TRN), a thin sheet of inhibitory neurons that surrounds the thalamus. The TRN inhibits the thalamus under control of the cortex, and is therefore known as the gatekeeper of the thalamus.

The authors focus on a gene called ErbB4, which is involved in the development and tuning of synapses onto inhibitory neurons. They first noticed that ErbB4 is specifically expressed in a particular type of neuron in the TRN called somatostatin, or SOM neurons. This was interesting because SOM neurons generally do not express ErbB4 elsewhere in the brain.

Intriguingly, both ErbB4 and its ligand Neuregulin1 are risk genes for schizophrenia. Since schizophrenia can involve deficits in sensory gating, implicating the TRN, the authors set out to discover what happens, both behaviorally and physiologically, when ErbB4 is deleted from SOM neurons in mice.

To assay the effect of this genetic manipulation, the authors developed a set of clever behavioral tasks. First, mice were trained on two basic alternative choice tasks. In the auditory version of the task, a high frequency sound indicated that the mouse should go right for a water reward, and a low frequency sound indicated the opposite. In the visual version, a light on the right indicated that the mouse should go right for reward, and vice versa. Both the normal wild-type and ErbB4 knockout mice learned this task equally well.

Next, the authors made life difficult for the mice by introducing distractors and misleading cues. In what they call the auditory-auditory task, the target cue is embedded in a sequence of distractor cues, which have a different timbre and are therefore distinguishable from the target cue. The mouse must ignore these distractors in order to successfully complete the task. In the visual-auditory task, mice that have been trained on both versions of the basic task are given both visual and auditory cues simultaneously. The mice must then learn to ignore the auditory cues, paying attention only to the visual ones.

Remarkably, the authors discovered that the knockout mice performed better on the auditory-auditory task, but worse on the visual-auditory task. The results were the same for a visual-visual task (a task with visual distractors) and an auditory-visual task (in which only the auditory cues are relevant). The authors conclude that mice lacking ErbB4 in somatostatin neurons in the TRN are better at filtering out distractors of the same modality, but worse at suppressing stimuli of the opposite modality that were once salient but now must be ignored.

The authors call this latter result a deficit in modality-switching, but a better experiment to assess modality-switching would be to train mice on only the basic auditory task before introducing them to the visual-auditory task. That way, the mice would truly have to switch their attention from one modality to the other, rather than suppress attention to one modality but not the other.

In an impressive set of physiological experiments, the authors demonstrate that this behavioral result is caused by the strengthening of a single type of synapse, from excitatory cells in the cortex onto somatostatin cells in the TRN.  They then propose a circuit model that explains their behavioral results. This model involves connections within the TRN between neurons of different modalities, the existence of which is controversial in the field. The validity of this model therefore awaits further study.

This is an excellent paper that combines genetics, electrophysiology and behavior to shed light on a neural circuit that is relevant to human disease. Although the final circuit model is tentative and relies on unproven anatomy, the paper convincingly demonstrates that changing the strength of a single class of synapse, within a single brain structure, can have profound effects on the behavior of an animal. Increasing the strength of this synapse improves within-modality suppression of distractors at the expense of cross-modality suppression of previously relevant cues. This furthers our understanding of how genes tune neural circuits to balance the many behavioral demands experienced by an animal.


Ahrens, et al, ErbB4 Regulation of a Thalamic Reticular Nucleus Circuit for Sensory Selection, Nature Neuroscience, Volume 18, No 1, January 2015