Motor learning involves neural circuit modifications in the cerebellar cortex, likely through re-weighting of parallel fiber inputs onto Purkinje cells (PCs). Climbing fibers instruct these synaptic modifications when they excite PCs in conjunction with parallel fiber activity, a pairing that enhances climbing fiber-evoked Ca2+ signaling in PC dendrites. In vivo, climbing fibers spike continuously, including during movements when parallel fibers are simultaneously conveying sensorimotor information to PCs. Whether parallel fiber activity enhances climbing fiber Ca2+ signaling during motor behaviors is unknown. In mice, we found that inhibitory molecular layer interneurons (MLIs), activated by parallel fibers during practiced movements, suppressed parallel fiber enhancement of climbing fiber Ca2+ signaling in PCs. Similar results were obtained in acute slices for brief parallel fiber stimuli. Interestingly, more prolonged parallel fiber excitation revealed latent supralinear Ca2+ signaling. Therefore, the balance of parallel fiber and MLI input onto PCs regulates concomitant climbing fiber Ca2+ signaling.

Understanding the activity of the mammalian brain requires an integrative knowledge of circuits at distinct scales, ranging from ion channel gating to circuit connectomics. Computational models are regularly employed to understand how multiple parameters contribute synergistically to circuit behavior. However, traditional models of anatomically and biophysically realistic neurons are computationally demanding, especially when scaled to model local circuits. To overcome this limitation, we trained several artificial neural network (ANN) architectures to model the activity of realistic multicompartmental cortical neurons. We identified an ANN architecture that accurately predicted subthreshold activity and action potential firing. The ANN could correctly generalize to previously unobserved synaptic input, including in models containing nonlinear dendritic properties. When scaled, processing times were orders of magnitude faster compared with traditional approaches, allowing for rapid parameter-space mapping in a circuit model of Rett syndrome. Thus, we present a novel ANN approach allowing for rapid, detailed network experiments using inexpensive and commonly available computational resources.

Proteomic profiling of brain cell types using isolation-based strategies pose limitations in resolving cellular phenotypes representative of their native state. We describe a mouse line for cell type-specific expression of biotin ligase TurboID, for in vivo biotinylation of proteins. Using adenoviral and transgenic approaches to label neurons, we show robust protein biotinylation in neuronal soma and axons throughout the brain, allowing quantitation of over 2000 neuron-derived proteins spanning synaptic proteins, transporters, ion channels and disease-relevant druggable targets. Next, we contrast Camk2a-neuron and Aldh1l1-astrocyte proteomes and identify brain region-specific proteomic differences within both cell types, some of which might potentially underlie the selective vulnerability to neurological diseases. Leveraging the cellular specificity of proteomic labeling, we apply an antibody-based approach to uncover differences in neuron and astrocyte-derived signaling phospho-proteins and cytokines. This approach will facilitate the characterization of cell-type specific proteomes in a diverse number of tissues under both physiological and pathological states.

The signals in cerebellar Purkinje cells sufficient to instruct motor learning have not been systematically determined. Therefore, we applied optogenetics in mice to autonomously excite Purkinje cells and measured the effect of this activity on plasticity induction and adaptive behavior. Ex vivo, excitation of channelrhodopsin-2-expressing Purkinje cells elicits dendritic Ca2+ transients with high-intensity stimuli initiating dendritic spiking that additionally contributes to the Ca2+ response. Channelrhodopsin-2-evoked Ca2+ transients potentiate co-active parallel fiber synapses; depression occurs when Ca2+ responses were enhanced by dendritic spiking. In vivo, optogenetic Purkinje cell activation drives an adaptive decrease in vestibulo-ocular reflex gain when vestibular stimuli are paired with relatively small-magnitude Purkinje cell Ca2+ responses. In contrast, pairing with large-magnitude Ca2+ responses increases vestibulo-ocular reflex gain. Optogenetically induced plasticity and motor adaptation are dependent on endocannabinoid signaling, indicating engagement of this pathway downstream of Purkinje cell Ca2+ elevation. Our results establish a causal relationship among Purkinje cell Ca2+ signal size, opposite-polarity plasticity induction, and bidirectional motor learning.

Synapse loss is associated with cognitive decline in Alzheimer’s disease, and owing to their plastic nature, synapses are an ideal target for therapeutic intervention. Oligomeric amyloid beta around amyloid plaques is known to contribute to synapse loss in mouse models and is associated with synapse loss in human Alzheimer’s disease brain tissue, but the mechanisms leading from Aβ to synapse loss remain unclear. Recent data suggest that the fast-activating and -inactivating voltage-gated potassium channel subtype 3.4 (Kv3.4) may play a role in Aβ-mediated neurotoxicity. Here, we tested whether this channel could also be involved in Aβ synaptotoxicity. Using adeno-associated virus and clustered regularly interspaced short palindromic repeats technology, we reduced Kv3.4 expression in neurons of the somatosensory cortex of APP/PS1 mice. These mice express human familial Alzheimer’s disease-associated mutations in amyloid precursor protein and presenilin-1 and develop amyloid plaques and plaque-associated synapse loss similar to that observed in Alzheimer’s disease brain. We observe that reducing Kv3.4 levels ameliorates dendritic spine loss and changes spine morphology compared to control virus. In support of translational relevance, Kv3.4 protein was observed in human Alzheimer’s disease and control brain and is associated with synapses in human induced pluripotent stem cell–derived cortical neurons. We also noted morphological changes in induced pluripotent stem cell neurones challenged with human Alzheimer’s disease-derived brain homogenate containing Aβ but, in this in vitro model, total mRNA levels of Kv3.4 were found to be reduced, perhaps as an early compensatory mechanism for Aβ-induced damage. Overall, our results suggest that approaches to reduce Kv3.4 expression and/or function in the Alzheimer’s disease brain could be protective against Aβ-induced synaptic alterations.

Eukaryotic cells maintain proteostasis through mechanisms that require cytoplasmic and mitochondrial translation. Genetic defects affecting cytoplasmic translation perturb synapse development, neurotransmission, and are causative of neurodevelopmental disorders, such as Fragile X syndrome. In contrast, there is little indication that mitochondrial proteostasis, either in the form of mitochondrial protein translation and/or degradation, is required for synapse development and function. Here we focus on two genes deleted in a recurrent copy number variation causing neurodevelopmental disorders, the 22q11.2 microdeletion syndrome. We demonstrate that SLC25A1 and MRPL40, two genes present in the microdeleted segment and whose products localize to mitochondria, interact and are necessary for mitochondrial ribosomal integrity and proteostasis. Our Drosophila studies show that mitochondrial ribosome function is necessary for synapse neurodevelopment, function, and behavior. We propose that mitochondrial proteostasis perturbations, either by genetic or environmental factors, are a pathogenic mechanism for neurodevelopmental disorders.

A common electrophysiology technique used in neuroscience is patch clamp: a method in which a glass pipette electrode facilitates single cell electrical recordings from neurons. Typically, patch clamp is done manually in which an electrophysiologist views a brain slice under a microscope, visually selects a neuron to patch, and moves the pipette into close proximity to the cell to break through and seal its membrane. While recent advances in the field of patch clamping have enabled partial automation, the task of detecting a healthy neuronal soma in acute brain tissue slices is still a critical step that is commonly done manually, often presenting challenges for novices in electrophysiology. To overcome this obstacle and progress towards full automation of patch clamp, we combined the differential interference microscopy optical technique with an object detection-based convolutional neural network (CNN) to detect healthy neurons in acute slice. Utilizing the YOLOv3 convolutional neural network architecture, we achieved a 98% reduction in training times to 18 min, compared to previously published attempts. We also compared networks trained on unaltered and enhanced images, achieving up to 77% and 72% mean average precision, respectively. This novel, deep learning-based method accomplishes automated neuronal detection in brain slice at 18 frames per second with a small data set of 1138 annotated neurons, rapid training time, and high precision. Lastly, we verified the health of the identified neurons with a patch clamp experiment where the average access resistance was 29.25 MΩ (n = 9). The addition of this technology during live-cell imaging for patch clamp experiments can not only improve manual patch clamping by reducing the neuroscience expertise required to select healthy cells, but also help achieve full automation of patch clamping by nominating cells without human assistance.