Rather than merely firing in a digital on–off fashion, vertebrate neurons may have an analogue aspect to their signalling too — a finding that will not surprise many who have worked on invertebrate neurons.
Much of what we know about electrical signalling in the brain comes from extracellular recordings that detect when a neuron is firing action potentials. These recordings do not, however, provide continuous monitoring of the fluctuations of membrane potential, and do not capture sub-threshold changes in membrane potential such as those caused by individual synaptic events. The prevalence of extracellular recordings in the literature has contributed to a collective consciousness in which the action potential or ‘spike’ is viewed as an invariant, all-or-nothing stereotyped event that occurs once a threshold membrane potential is reached. This ‘digital’ signal carries information from the neuronal cell body, the soma, down the axon to presynaptic terminals, where it evokes the release of neurotransmitter to excite or inhibit the next neuron. In this simple framework, the spike frequency might influence the amount of transmitter release by mechanisms such as facilitation and depression1, but sub-threshold events occurring in the soma and the dendrites (the projections that receive inputs from other neurons) are thought to be too far away to influence transmitter release from the axonal terminals.
Alle and Geiger writing in Science2, and Shu et al. (this issue, page 761)3 now show, however, that the release of neurotransmitter from axon terminals of certain vertebrate neurons is influenced by the somatic membrane potential. So, in terms of neuronal signalling, the axon terminals and the soma are electrically much closer than many would have assumed (Fig. 1). The salient finding in both papers, from which all of the rest of the results follow, is that the length constant, λ, of axons is surprisingly long, at about 420–450 µm, in two types of neuron: hippocampal mossy fibres from rats2 and layer 5 pyramidal cells from the prefrontal cortex of ferrets3. λ is the distance over which a voltage change imposed at one site will drop to approximately 37% of its initial value4. Although rapid changes in membrane potential are more attenuated by distance than are slow signals, when two regions of a neuron are less than λ apart, they are commonly considered to be electrically ‘close’. Put another way, at a distance less than λ, changes in membrane potential at one place will appreciably alter the membrane potential at the other. In the case of the mossy fibres, λ was calculated using the distance between recordings made from boutons (presumably presynaptic release sites) and the soma2. In the case of the prefrontal cortical neurons, simultaneous recordings were made from the soma and axon at known distances apart3.
How surprising is it that λ is so long in these neurons? Some readers will undoubtedly be taken aback to discover that vertebrate neurons can be so electrically compact, at least for slow signals. But those of us who work with large neurons in invertebrates already know of cases in which somatic voltages influence the axon more than a millimetre away5. In the vertebrate hippocampal2 and layer 5 cortical neurons3 the consequence of λ being so long is that slow depolarizations (positive-going changes in membrane potential that bring the neuron closer to threshold) of the membrane at the somatic and dendritic regions are transmitted to the axon terminals, and can influence the release of transmitter.
Shu et al.3 show that depolarization of the soma can increase the amount of spike-mediated release of neurotransmitter from the pre-synaptic terminals by about 30% per 10 mV of somatic depolarization. Is this change likely to be important for the function of the circuits containing these neurons? This will depend on many factors, but it is fair to say that even small changes in synaptic strength can have significant influences on circuit performance.
In the work by Shu et al.3 somatic depolarization was associated with significant spike broadening, that is the return to resting potential occurred more slowly, in both the soma and the axon. Presumably this spike broadening may contribute to the enhanced transmitter release, as occurs elsewhere6,7. In this study3, the authors recorded ongoing barrages of synaptic potentials resulting from spontaneous network activity. The depolarizations associated with these synaptic barrages were seen in both somatic and axonal recordings, demonstrating that signals of the amplitude and time-course of normal synaptic inputs could influence axonal membrane potential.
Alle and Geiger2 saw no evidence of spike broadening in their recordings. But they did show that somatic depolarizations that mimic those recorded in vivo during theta oscillations (an ongoing brain rhythm) can substantially increase the amplitudes of the postsynaptic currents evoked by action potentials.
Many types of invertebrate neuron release neurotransmitter both in response to action potentials and as a graded function of membrane potential8,9,10,11. There is, however, an important difference between the results in the two studies highlighted here2,3 and the classic graded transmission well characterized in the retina and in many invertebrate neurons12,13,14,15,16. In classic graded transmission, action potentials are not necessary for transmitter release because the threshold for transmitter release is often close to the resting potential. Therefore, synaptic release follows the analogue fluctuations of the presynaptic neuron's membrane potential. In contrast, the results of Alle and Geiger2 and of Shu et al.3 show no indication that transmitter release occurs independently of action potentials, but only that the somatic membrane potential can influence how much transmitter is released by an action potential.
Over the years, I have watched the characteristics ascribed to vertebrate neurons slowly evolve to exhibit many of the attributes so clearly demonstrable in invertebrate neurons. I now find it gratifying that vertebrate neurons found in the hippocampus and cortex, brain regions well known for their importance in memory and cognition, may also avail themselves of some of the benefits of analogue processing in addition to spike-mediated release. My guess is that this work2,3 will trigger numerous studies to determine how transmitter release from spiking neurons in the vertebrate brain is influenced by slow changes in membrane potential that result from neuromodulation and ongoing activity. In this way, we will come to understand how neural circuits produce variable outputs as a function of behavioural state and mood.
About this article
Nature Reviews Neuroscience (2013)