Among your brain's many responsibilities, one of the most important is distinguishing between external and self-generated signals. The sound of one's own laughter might closely resemble the shrieking of an irascible badger, but the two auditory stimuli require wildly different responses. One way that the brain might differentiate between these two scenarios is by actively
The best evidence for this important theory comes from a bizarre and handsome animal: the weakly electric fish. In order to navigate through muddy streams, Mormyrid fish use a specialized organ to produce a weak electrical pulse and monitor deformations of the resulting current flow with passive electroreceptors in the skin, thus forming an electrical ‘image’ of their environment. Previous studies have found that a corollary discharge signal from the electric organ is subtracted from sensory inputs within a specific circuit of the fish brain called the electrosensory lobe. However, in addition to predicting the stereotyped discharge of the electric organ, the brain must also take into account the seemingly infinite number of different movements produced by a frisky fish. Is there a different pattern of corollary discharge signals for each of these possible movements? If so, how are these myriad combinations integrated with electrosensory information in the brain?
A recent paper by Tim Requarth and Nate Sawtell at Columbia University, USA, addressed these questions by measuring the responses of neurons in the electrosensory lobe to movements of the tail. The authors evoked rhythmic swimming movements within paralyzed fish by electrically stimulating locomotor circuits in the brain. They then recorded extracellularly from mossy fibers, neurons that convey motor signals from the spinal cord to the electrosensory lobe. They discovered that some mossy fibers encode unique movement parameters, such as the frequency, amplitude or direction of tail wagging. Thus, different mossy fibers carry corollary discharge signals related to specific aspects of the fish's self-movement.
Requarth and Sawtell next investigated the convergence of motor and electrosensory signals within the principal neurons of the fish's electrosensory lobe. Mossy fibers connect to tiny neurons called granule cells, which subsequently route self-movement signals to principal neurons. When granule cell signals are repeatedly paired with electrosensory input, the strength of the synapse between the granule cell and the principal neuron decreases. This allows the principal neuron to subtract predictable components of the sensory response that are due to the fish's own behavior. The authors observed that repeated pairing of a tail movement with coincident electric organ discharge led to a highly specific depression of granule cell inputs. Furthermore, they found that each principal neuron was capable of subtracting multiple unique motor signals from a given sensory event.
These findings provide a tantalizing hint of how the brain might cope with the extreme dimensionality of possible motor behaviors. Because a single principal neuron in the electrosensory lobe receives synaptic input from many granule cells (~20,000), each principal neuron could potentially subtract gazillions of different movement combinations from a given sensory signal. The basic circuit structure of the Mormyrid electrosensory lobe is much like that of the cerebellum, a cauliflower-like structure at the rear of the mammalian brain, which is thought to play a similar role in predicting the sensory consequences of motor commands. Therefore, insights from the electric fish might help us understand brain circuits in furrier animals, such as humans or badgers.
