Analogue brains

Digital computers and mammalian brains share not only a talent for computation but also a method that favors absolutes. The ones and zeros of a computer are analogous to the all-or-none character of action potentials in the brain. Neurons sum all incoming depolarizing signals until, at a threshold level of depolarization, an action potential fires along the axon. Now, however, Henrik Alle and Jörg Geiger (Max Planck, Frankfurt, Germany) show that the rat brain also uses graded signals for communication.

This analogue or subthreshold processing has been seen in invertebrates, but “in the central mammalian brain it was never shown,” says Geiger. The presumption was that the all-or-none action potentials were the whole story. “To be honest,” says Geiger, “I don't know how this kind of thinking developed.”

Alle and Geiger looked in the hippocampus, the center of memory and learning, using tissue slices. They introduced subthreshold signals in the cell body and saw that these reached synapses up to 1 mm away. If the subthreshold signals reached the synapse just before an action potential arrived, the response in the postsynaptic neuron was increased.

Although the information in the subthreshold signal can only be communicated in combination with an action potential, it may have great power. Most synapses in the brain are close to their cell bodies— within range of the subthreshold signals. And so-called theta oscillations are a potentially crucial form of subthreshold signaling. These oscillations are thought to act as a kind of clock in the brain, helping to “bind” different sensory inputs into a single experience.

Geiger now plans to look for subthreshold signals generated by receptors in the brain, and to test whether individual axons are carrying out analogue computations. The molecular mechanisms that transduce the signal are also a mystery. “Classically one would expect it to be calcium, but we have evidence against that,” says Geiger. “This is completely new territory.”

Reference: