Biology subjects

Stuart, G.

Publications and source records attributed to Stuart, G..

4 recordsLinked to original sources

bioRxiv2024

Biologically Realistic Computational Primitives of Neocortex Implemented on Neuromorphic Hardware Improve Vision Transformer Performance

Understanding the computational principles of the brain and replicating them on neuromorphic hardware and modern deep learning architectures is crucial for advancing neuro-inspired AI (NeuroAI). Here, we develop an experimentally-constrained biophysical network model of neocortical circuit motifs, focusing on layers 2-3 of the primary visual cortex (V1). We investigate the role of four major cortical interneuron classes in a competitive-cooperative computational primitive and validate these circuit motifs implemented soft winner-take-all (sWTA) computation for gain modulation, signal restoration, and context-dependent multistability. Using a novel parameter mapping technique, we configured IBMs TrueNorth (TN) chip to implement sWTA computations, mirroring biological neural dynamics. Retrospectively, we observed a strong correspondence between the biophysical model and the TN hardware parameters, particularly in the roles of four key inhibitory neuron classes: Parvalbumin (feedforward inhibition), Somatostatin (feedback inhibition), VIP (disinhibition), and LAMP5 (gain normalization). Moreover, the sparse coupling of this sWTA motif was also able to simulate a two-state neural state machine on the TN chip, replicating working memory dynamics essential for cognitive tasks. Additionally, integrating the sWTA computation as a preprocessing layer in the Vision Transformer (ViT) enhanced its performance on the MNIST digit classification task, demonstrating improved generalization to previously unseen data and suggesting a mechanism akin to zero-shot learning. Our approach provides a framework for translating brain-inspired computations to neuromorphic hardware, with potential applications on platforms like Intels Loihi2 and IBMs Northpole. By integrating biophysically accurate models with neuromorphic hardware and advanced machine learning techniques, we offer a comprehensive roadmap for embedding neural computation into NeuroAI systems.

neuroscience↗

bioRxiv2024

Cellular and circuit mechanisms underlying binocular vision

How binocular visual information is combined at the level of single neurons in the brain is not fully understood. Here, we show in mice that callosal input from the opposite visual cortex (V1) plays a critical role in this process. In vivo we find this callosal projection carries ipsilateral eye information and synapses exclusively onto binocular neurons. Using the presence of callosal input to identify binocular neurons in vitro, at the cellular level we show that binocular neurons are less excitable than monocular neurons due to high expression of Kv1 potassium channels. At the circuit level we find that only monocular neurons send callosal projections to the opposite V1, whereas binocular neurons do not. Finally, using dual-colour optogenetics we show that most binocular and monocular neurons receive direct input from the thalamus. In summary, we describe distinct cellular and circuit mechanisms underlying processing of binocular visual information in mouse V1.

neuroscience↗

bioRxiv2020

Dendritic spikes in apical oblique dendrites of cortical layer 5 pyramidal neurons

Dendritic spikes in layer 5 pyramidal neurons (L5PNs) play a major role in cortical computation. While dendritic spikes have been studied extensively in apical and basal dendrites of L5PNs, whether oblique dendrites, which ramify in the input layers of the cortex, also generate dendritic spikes is unknown. Here we report the existence of dendritic spikes in apical oblique dendrites of L5PNs. In silico investigations indicate that oblique branch spikes are triggered by brief, low-frequency action potential (AP) trains (~40 Hz) and are characterized by a fast sodium spike followed by activation of voltage-gated calcium channels. In vitro experiments confirmed the existence of oblique branch spikes in L5PNs during brief AP trains at frequencies of around 60 Hz. Oblique branch spikes offer new insights into branch-specific computation in L5PNs and may be critical for sensory processing in the input layers of the cortex.

neuroscience↗

bioRxiv2018

Deleterious genetic variants in NOTCH1 are a major contributor to the incidence of non-syndromic Tetralogy of Fallot

AimsFamilial recurrence studies provide strong evidence for a genetic component to the predisposition to sporadic, non-syndromic Tetralogy of Fallot (TOF), the most common cyanotic congenital heart disease (CHD) phenotype. Rare genetic variants have been identified as important contributors to the risk of CHD, but relatively small numbers of TOF cases have been studied to date. Here, we use whole exome sequencing to assess the prevalence of rare, potentially deleterious variants in candidate genes previously associated with both syndromic and non-syndromic TOF, in the largest cohort of non-syndromic TOF patients reported to date.\n\nMethods & Results829 non-syndromic TOF patients underwent whole exome sequencing. A systematic review of the literature was conducted which revealed 77 genes in which mutations had been reported in patients with TOF. The presence of rare, deleterious variants in the 77 candidate genes was determined, defined by a minor allele frequency of [≤] 0.001 and scaled combined annotation-dependent depletion (CADD) score of [≥] 20. We found a clustering of heterozygous rare, deleterious variants in NOTCH1 (P=1.89E-15), DOCK6 (P=2.93E-07), MYOM2 (P= 7.35E-05), TTC37 (P=0.016), MESP1 (P=0.024) and TBX1 (P=0.039), after correcting for multiple testing. NOTCH1 was most frequently found to harbour deleterious variants. Changes were observed in 49 patients (6%; 95% confidence interval [CI]: 4.5% - 7.8%) and included six truncating/frameshift variants and forty missense variants. Sanger sequencing of the unaffected parents of thirteen cases identified five de novo variants. Variants were not confined to a single functional domain of the NOTCH1 protein but significant clustering of variants was evident in the EGF-like repeats (P=0.018). Three NOTCH1 missense variants (p.G200R, p.C607Y and de novo p.N1875S) were subjected to functional evaluation and showed a reduction in Jagged1 ligand-induced NOTCH signalling. p.C607Y, which exhibited the most significant reduction in signalling, also perturbed S1 cleavage of the NOTCH1 receptor in the Golgi.\n\nConclusionThe NOTCH1 locus is a frequent site of genetic variants predisposing to non-syndromic TOF with 6% of patients exhibiting rare, deleterious variants. Our data supports the polygenic origin of TOF and suggests larger studies may identify additional loci.

genetics↗