This Week in The Journal

The Role of Spike Rate in Synaptic Plasticity

Michael Graupner, Pascal Wallisch, and Srdjan Ostojic

When asked to describe the neural mechanisms underlying learning, most neuroscientists will describe spike-timing dependent plasticity: the connection between two neurons is strengthened when a presynaptic spike is followed immediately by a postsynaptic spike. This type of plasticity has been demonstrated in numerous experiments. But spike timing is only one driver of plasticity; other factors, such as firing rate, appear to be more important at some synapses. In fact, the relative importance of spike timing and spike rate for plasticity at most synaptic classes in vivo is unclear, because spiking patterns imposed in experimental studies typically do not resemble those occurring naturally. In particular, while most studies induce regular spiking in presynaptic neurons and impose fixed delays between presynaptic spikes and postsynaptic responses, spiking in vivo is highly variable in many neuronal populations.

To investigate the relative importance of spike timing and spike rate in the induction of synaptic plasticity by natural spike trains, Graupner et al. turned to computational models. They modeled two reciprocally connected neurons and used various rules to govern plasticity: in a standard spike-pair model, plasticity was determined solely by the delay between presynaptic and postsynaptic spikes; a similar model incorporated spike triplets; and in two additional models, plasticity was influenced by postsynaptic calcium levels, which were modulated linearly or nonlinearly.

The authors examined synaptic changes elicited when model neurons were subjected to spike pairs with fixed time lags, correlated spike trains, or irregular spike patterns recorded in cortex of awake monkeys. For spike pairs, precise timing was less important when spikes occurred irregularly than when they occurred regularly. More importantly, for all plasticity models, modulating firing rate without considering timing produced plasticity comparable to that produced by altering the proportion of correlated presynaptic and postsynaptic spikes. The effects of modulating spike correlations or rates depended on the initial firing rate of the model neurons.

These results demonstrate that modestly increasing spike rate is able to drive synaptic plasticity as well as generating precisely coupled presynaptic and postsynaptic spikes. The authors suggest that whether spike timing or rate is most important in a given neuronal population will depend on the neurons' baseline firing rate and variability.

Control of Feeding by Septal Neurons

Patrick Sweeney and Yunlei Yang

(see pages 11185–11195)

The drive to eat is influenced by numerous factors, including energy balance, reward, social context, and emotions. These influences converge on the lateral hypothalamic area, which in turn projects to forebrain, midbrain, and hindbrain areas that control motivation, arousal, locomotion, and autonomic function. Electrical stimulation of the lateral hypothalamic area is reinforcing (rodents will press a lever to receive such stimulation) and it causes voracious eating. These effects are reproduced by selectively activating GABAergic neurons in the lateral hypothalamus. In contrast, activation of glutamatergic neurons in this area reduce feeding even in food-deprived mice, and mice avoid locations associated with this stimulation (Stuber and Wise 2016 Nat Neurosci 19:198). How the activity of these neurons is modulated by inputs conveying information about energy balance, emotions, and food availability is only beginning to be revealed.

Optogenetic activation of septal GABAergic terminals (green) in the lateral hypothalamus induces inhibitory currents in local GABAergic neurons (red). See Sweeney and Yang for details.

The regulation of feeding by social and emotional states might be mediated in part by projections from the septum to the lateral hypothalamus. Because previous studies suggested that the septum regulates feeding, Sweeney and Yang expressed a stimulatory designer drug exclusively activated by designer receptors (DREADD) in GABAergic septal neurons. Activating these neurons reduced feeding. Similarly, optogenetic activation of septal GABAergic terminals in the lateral hypothalamus reduced food intake. Conversely, inhibiting septal GABAergic neurons with an inhibitory DREADD increased food intake.

Recordings in brain slices demonstrated that ∼20% of GABAergic neurons in the lateral hypothalamus received monosynaptic inputs from septal GABAergic neurons. Consistent with previous studies, chemogenetic activation of GABAergic neurons in the lateral hypothalamus increased food consumption. Importantly, however, when septal GABAergic neurons were stimulated optically while lateral hypothalamic GABAergic neurons were activated chemogenetically, effects on feeding were eliminated.

Together, these data suggest that GABAergic neurons in the septum reduce feeding at least partly by inhibiting GABAergic neurons in the lateral hypothalamus that normally promote feeding. Future work should identify afferents that activate feeding-related septal neurons. Such work might deepen our understanding of the neural mechanisms underlying stress-induced over- and under-eating, which might in turn lead to treatments for certain eating disorders.