Brains process external information rapidly at a sub-second time scale, which is set by the dynamic electrophysiological properties of neurons and the fast communication within neuronal populations. This fast neural processing is complemented by the so-called neuromodulatory systems (involving certain class of neurotransmitters such as dopamine and serotonin). Neuromodulation has generally been thought to occur at slower time scales, for example during periods of alertness following some salient event or over circadian periods for sleep-wake modulation. The fast and slow systems working together allow the brain to not only react rapidly to external stimuli, but also assign context or meaning to these stimuli. In our recent study, appearing in Nature Neuroscience, we set out to understand how a particular neuromodulatory system involving serotonin influences information processing in a sensory system. What we found was unexpected and exciting.
Serotonin is a chemical that has been linked to high level cognitive features such as depression, aggression and mood. Although we are far from understanding the neuronal architecture underlying any of these effects, it is generally theorized that the serotonin system affects neural processing by slowly altering the properties of circuit elements, the neurons and synapses. In mammals, serotonin is secreted by neurons located in the raphe nuclei, which send their axons widely throughout the brain, including very dense projections to the early stages of olfactory system. This fact, combined with the importance of olfaction for mice, prompted us to examine the involvement of the serotonin system in odor processing.
Our experiments were enabled by the explosive advances in neuroscience techniques, including optogenetics (which allowed us to selectively activate specific neurons and axons with light) and optical reporters of activity (genetically-encoded calcium indicators that transduce neural activity to light). We used multiphoton microscopy to look at the activity of two different populations of output neurons (mitral and tufted cells) in the olfactory bulb, the first odor processing stage in vertebrates. To our surprise, we found that even brief activation of raphe neurons caused immediate excitation of mitral and tufted cells. An even greater surprise was in store when we complemented our whole animal experiments with mechanistic studies in ex vivo brain slices; in addition to releasing serotonin, raphe neurons also released a fast excitatory neurotransmitter, glutamate. In fact, glutamate mediates much of the excitation of mitral and tufted cells in our experiments, with serotonin release likely requiring more intense activity in raphe neurons.
Sensory systems are not only required to detect external stimuli (odors in the case of the olfactory system), but they also need to make distinctions between different stimuli. We asked how the activation of raphe neurons modulates these functions of the olfactory system. Uncharacteristically for a neuromodulatory system, qualitatively distinct effects were seen in the two types of olfactory bulb neurons: activating raphe neurons enhanced the response of tufted cells to odors, but bi-directionally modulated the odor response of mitral cells. A quantitative analysis of the population coding of odors revealed that raphe activation makes tufted cells more sensitive to detecting odors and the mitral cells better at discriminating different odors.
Overall, our study indicates that a “neuromodulatory” system, traditionally considered to have slow actions, can actually be part of fast ongoing information processing by releasing multiple types of neurotransmitters. Further, these modulatory neurons need not have monolithic effects, but can influence different channels of information processing in distinct ways. Conceptually, our study blurs the distinction between neuromodulation and computation itself.This research was supported by grants from the NIH, a seed grant from the Harvard Brain Initiative and fellowships from the NSF and the Sackler Foundation.
Read more in Nature Neuroscience or download PDF
Read more in News and Views or download PDF