Protein engineering and experiment design: JSM, LLL Hippocampal slice imaging: TPJ, DAR Visual cortex volume imaging: KP, ON Mouse epilepsy model: YS, VM, ML, DMK Mitochondria experiments: ELK, NJL Zebrafish: TK, MBA Abstract (150 words)Current techniques for monitoring GABA, the primary inhibitory neurotransmitter in vertebrates, cannot follow ephemeral transients in intact neural circuits. We applied the design principles used to create iGluSnFR, a fluorescent reporter of synaptic glutamate, to develop a GABA sensor using a protein derived from a previously unsequenced Pseudomonas fluorescens strain. Structure-guided mutagenesis and library screening led to a usable iGABASnFR (∆F/Fmax ~ 2.5, Kd ~ 9 µM, good specificity, adequate kinetics). iGABASnFR is genetically encoded, detects single action potential-evoked GABA release events in culture, and produces readily detectable fluorescence increases in vivo in mice and zebrafish. iGABASnFR enabled tracking of: (1) mitochondrial GABA content and its modulation by an anticonvulsant; (2) swimming-evoked GABAergic transmission in zebrafish cerebellum; (3) GABA release events during inter-ictal spikes and seizures in awake mice; and (4) GABAergic tone decreases during isoflurane anesthesia. iGABASnFR will permit high spatiotemporal resolution of GABA signaling in intact preparations.

Although cortical interneurons are apparently well-placed to suppress seizures, several recent reports have highlighted a paradoxical role of parvalbumin-positive perisomatic-targeting (PV+) interneurons in ictogenesis. Here, we use an acute in vivo model of focal cortical seizures in awake behaving mice, together with closed-loop optogenetic manipulation of PV+ interneurons, to investigate their function during seizures. We show that photo-depolarization of PV+ interneurons rapidly switches from an anti-ictal to a pro-ictal effect within a few seconds of seizure initiation. The pro-ictal effect of delayed photostimulation of PV+ interneurons was not shared with dendrite-targeting somatostatin-positive (SOM+)interneurons. We also show that this switch can be prevented by overexpression of the neuronal potassium-chloride co-transporter KCC2 in principal cortical neurons. These results suggest that strategies aimed at improving the ability of principal neurons to maintain intracellular chloride levels in the face of excessive network activity can prevent interneurons from contributing to seizure perpetuation.

See Karakaya and Wirth (doi:10.1093/brain/awz273) for a scientific commentary on this article. Neurofascin (NFASC) isoforms are immunoglobulin cell adhesion molecules involved in node of Ranvier assembly. Efthymiou et al. identify biallelic NFASC variants in ten unrelated patients with a neurodevelopmental disorder characterized by variable degrees of central and peripheral involvement. Abnormal expression of Nfasc155 is accompanied by severe loss of myelinated fibres.

The units in artificial neural networks (ANNs) can be thought of as abstractions of biological neurons, and ANNs are increasingly used in neuroscience research. However, there are many important differences between ANN units and real neurons. One of the most notable is the absence of Dale's principle, which ensures that biological neurons are either exclusively excitatory or inhibitory. Dale's principle is typically left out of ANNs because its inclusion impairs learning. This is problematic, because one of the great advantages of ANNs for neuroscience research is their ability to learn complicated, realistic tasks. Here, by taking inspiration from feedforward inhibitory interneurons in the brain we show that we can develop ANNs with separate populations of excitatory and inhibitory units that learn just as well as standard ANNs. We call these networks Dale's ANNs (DANNs). We present two insights that enable DANNs to learn well: (1) DANNs are related to normalization schemes, and can be initialized such that the inhibition centres and standardizes the excitatory activity, (2) updates to inhibitory neuron parameters should be scaled using corrections based on the Fisher Information matrix. These results demonstrate how ANNs that respect Dale's principle can be built without sacrificing learning performance, which is important for future work using ANNs as models of the brain. The results also may have interesting implications for how inhibitory plasticity in the real brain operates.

Fast inhibitory signaling in the brain has conventionally been considered to be predominantly mediated by the vesicular release of GABA from presynaptic terminals onto postsynaptic GABAA receptors.1 Transient opening of such receptors results in a brief increase in postsynaptic permeability to Cl–, generating an inhibitory postsynaptic potential (IPSP) that reduces the probability of firing of the neuron. However, there is abundant evidence that GABA can also act relatively far from its site of release, and this, together with several other discoveries in the last two decades, has contributed to a reappraisal of the roles of GABAA receptors in modulating neuronal and circuit excitability.1