bioRxiv Subject Collection: Neuroscience's Journal

Network-based amyloid-beta pathology predicts subsequent cognitive decline in cognitively normal older adults
The deposition of amyloid-beta protein in the human brain is a hallmark of Alzheimer's disease and is related to cognitive decline. However, the relationship between early amyloid-beta deposition and future cognitive impairment remains poorly understood, particularly concerning its spatial distribution and network-level effects. Here, we employed a cross-validated machine learning approach and investigated whether integrating subject-specific brain connectome information with amyloid-beta burden measures improves predictive validity for subsequent cognitive decline. Baseline regional amyloid-beta pathology measures from positron emission tomography (PET) imaging predicted prospective cognitive decline. Incorporating structural connectome, but not functional connectome, information into the amyloid-beta measures improved predictive performance. We further identified a neuropathological signature pattern linked to future cognitive decline, which was validated in an independent cohort. These findings advance our understanding of how amyloid-beta pathology relates to brain networks and highlight the potential of network-based metrics for amyloid-beta-PET imaging to identify individuals at higher risk of cognitive decline.

Investigating cerebellar control in slow gait adaptations: Insights from predictive simulations of split-belt walking
During split-belt treadmill walking, neurotypical humans exhibit slow adaptations, characterized by a gradual decrease in step length asymmetry, whereas individuals with cerebellar damage do not show these motor adaptations. We used a neuromusculoskeletal model to better understand individual aspects of the underlying neural control. Specifically, we extended a spinal reflex model by adding a supraspinal layer, representing the cerebellum and its main function of error-driven motor adaptation. The cerebellum, based on the mismatch between an internal prediction and the actual motor outcome, can modulate spinal motor commands within the simulation. Using this model, we investigated the effect of an isolated adaptation of gait timing parameters, in our case the beginning of the liftoff phase. We created 80 s predictive simulations of the model walking on a split-belt treadmill with a 2:1 belt-speed ratio, and evaluated the results by comparing spatiotemporal parameters and kinematics with literature. The simulations exhibited adaptation patterns similar to those observed in human experiments. Specifically, the step length symmetry decreased from an initial asymmetric level toward the baseline, driven primarily by adaptations in the fast step length, while the individual joint kinematics remained similar. The adaptations affected the spatial and temporal domains, represented by a change in the center of oscillation difference and limb phasing. Our findings suggest that reflex gains do not necessarily need to be adapted to achieve changes in step length asymmetry and that, unlike what had been inferred from experiments, the same neural mechanism might account for adaptations in the spatial and temporal domains at different rates. Our simulations demonstrated distinct adaptation patterns corresponding to slow and fast learning behaviors, as reported in the literature, through modifications of a single cerebellar parameter, the adaptation rate. The framework can be extended to test different hypotheses about motor control and adaptations during continuous perturbation tasks.

Data-driven synapse classification reveals a logic of glutamate receptor composition
The rich diversity of synapses facilitates the capacity of neural circuits to transmit, process and store information. Here, we used multiplex super-resolution proteometric imaging through array tomography to define features of single synapses in the adult mouse neocortex. We find that glutamatergic synapses cluster into subclasses that parallel the distinct biochemical and functional categories of receptor subunits: GluA1/4, GluA2/3 and GluN1/GluN2B. Two of these subclasses align with physiological expectations based on synaptic plasticity: large AMPAR-rich synapses may represent potentiated synapses, whereas small NMDAR-rich synapses suggest "silent" synapses. The NMDA receptor content of large synapses correlates with spine neck diameter, and thus the potential for coupling to the parent dendrite. Conjugate array tomography's rigorous registration of immunofluorescence with electron microscopy provides validation for future super-resolution imaging studies in other systems. No barriers prevent generalization of this approach to other species, laying a foundation for future studies of human disorders and therapeutics.

Single neuron diversity supports area functional specialization along the visual cortical pathways
Humans and other primates have specialized visual pathways composed of interconnected cortical areas. The input area V1 contains neurons that encode basic visual features, whereas downstream in the lateral prefrontal cortex (LPFC) neurons acquire tuning for novel complex feature associations. It has been assumed that each cortical area is composed of repeatable neuronal subtypes, and variations in synaptic strength and connectivity patterns underlie functional specialization. Here we test the hypothesis that diversity in the intrinsic make-up of single neurons contributes to area specialization along the visual pathways. We measured morphological and electrophysiological properties of single neurons in areas V1 and LPFC of marmosets. Excitatory neurons in LPFC were larger, less excitable, and fired broader spikes than V1 neurons. Some inhibitory fast spiking interneurons in the LPFC had longer axons and fired spikes with longer latencies and a more depolarized action potential trough than in V1. Intrinsic bursting was found in subpopulations of both excitatory and inhibitory LPFC but not V1 neurons. The latter may favour temporal summation of spikes and therefore enhanced synaptic plasticity in LPFC relative to V1. Our results show that specialization within the primate visual system permeates the most basic processing level, the single neuron.

Human Hippocampal Theta Oscillations Organise Distance to Goal Coding
The rodent hippocampal local field potential is dominated by 6-12 Hz theta oscillations during active behaviour that are strongly implicated in spatial coding and memory function across species. Invasive electrophysiology in both rodents and humans has shown increases in hippocampal theta power immediately before the onset of translational movement that persists throughout subsequent motion, and the magnitude of this increase correlates with the distance subsequently travelled. Using non-invasive magnetoencephalography (MEG) and an abstract navigation task, we observed increased theta power during both spatial planning and subsequent navigation. Importantly, theta power in the right hippocampus covaried with subsequent path distance during planning, only when participants were aware of the distance to their goal. During subsequent navigation, hippocampal theta power decreased dynamically as participants approached the goal, only when they were aware of how far they still needed to travel. In addition, theta phase during navigation modulated 70-140 Hz fast gamma amplitude in the entorhinal cortex while traversing novel paths; and 30-70 Hz slow gamma amplitude in the right hippocampus while traversing previously experienced paths. In both cases, theta-gamma phase-amplitude coupling increased with proximity to the goal during navigation, consistent with the hypothesis that sequences of upcoming locations were represented by gamma bursts occurring at successive theta phases. In sum, these findings are consistent with the proposed role of hippocampal theta oscillation in flexible planning and goal-directed spatial navigation across mammalian species.

GABA, Glutamate dynamics and BOLD observed during cognitive processing in psychosis patients with hallucinatory traits
The perception of a voice in the absence of an external auditory source - an auditory verbal hallucination - is a characteristic symptom of schizophrenia. To better understand this phenomenon requires integration of findings across behavioural, functional, and neurochemical levels. We address this with a locally adapted MEGA-PRESS sequence incorporating interleaved unsuppressed water acquisitions, allowing concurrent assessment of behaviour, blood-oxygenation-level-dependent (BOLD) functional changes, Glutamate+Glutamine (Glx), and GABA, synchronised with a cognitive (flanker) task. We acquired data from the anterior cingulate cortex (ACC) of 51 patients with psychosis (predominantly schizophrenia spectrum disorder) and hallucinations, matched to healthy controls. Consistent with the notion of an excitatory/inhibitory imbalance, we hypothesized differential effects for Glx and GABA between groups, and aberrant dynamics in response to task. Results showed impaired task performance, lower baseline Glx and positive association between Glx and BOLD in patients, contrasting a negative correlation in healthy controls. Task-related increases in Glx were observed in both groups, with no significant difference between groups. No significant effects were observed for GABA. These findings suggest that a putative excitatory/inhibitory imbalance affecting inhibitory control in the ACC is primarily observed as tonic, baseline glutamate differences, rather than GABAergic effects or aberrant dynamics in relation to a task.

Perinatal serotonin signalling dynamically influences the development of cortical GABAergic circuits with consequences for lifelong sensory encoding
Serotonin plays a prominent role in neurodevelopment, regulating processes from cell division to synaptic connectivity1,2. Clinical studies suggest that alterations in serotonin signalling such as genetic polymorphisms3-5 or antidepressant exposure during pregnancy6,7 are risk factors for neurodevelopmental disorders. However, an understanding of how dysfunctional neuromodulation alters systems level activity over neocortical development is lacking. Here, we use a longitudinal imaging approach to investigate how genetics, pharmacology, and aversive experience disrupt state-dependent serotonin signalling with pathological consequences for sensory processing. We find that all three factors lead to increased neocortical serotonin levels during the initial postnatal period. Genetic deletion of the serotonin transporter or antidepressant dosing results in a switch from hypo- to hyper-cortical activity that arises as a consequence of altered cortical GABAergic microcircuitry. However, the trajectories of these manipulations differ with postnatal exposure to antidepressants having effects on adult sensory encoding. The latter is not seen in the genetic model despite a similar early phenotype, and a distinct influence of maternal genotype on the development of supragranular layers. These results reveal the dynamics and critical nature of serotonin signalling during perinatal life; pharmacological targeting of which can have profound life-long consequences for cognitive development of the offspring.

LMX1B missense-perturbation of regulatory element footprints disrupts serotonergic forebrain axon arborization
Pathogenic coding mutations are prevalent in human neuronal transcription factors (TFs) but how they disrupt development is poorly understood. Lmx1b is a master transcriptional regulator of postmitotic Pet1 neurons that give rise to mature serotonin (5-HT) neurons; over two hundred pathogenic heterozygous mutations have been discovered in human LMX1B, yet their impact on brain development has not been investigated. Here, we developed mouse models with different LMX1B DNA-binding missense mutations. Missense heterozygosity broadly altered Pet1 neuron transcriptomes, but expression changes converged on axon and synapse genes. Missense heterozygosity effected highly specific deficits in the postnatal maturation of forebrain serotonin axon arbors, primarily in the hippocampus and motor cortex, which was associated with spatial memory defects. Digital genomic footprinting (DGF) revealed that missense heterozygosity caused complete loss of Lmx1b motif protection and chromatin accessibility at sites enriched for a distal active enhancer/active promoter histone signature and homeodomain binding motifs; at other bound Lmx1b motifs, varying levels of losses, gains or no change in motif binding and accessibility were found. The spectrum of footprint changes was strongly associated with synapse and axon genes. Further, Lmx1b missense heterozygosity caused wide disruption of Lmx1b-dependent GRNs comprising diverse TFs expressed in Pet1 neurons. These findings reveal an unanticipated continuum of Lmx1b missense-forced perturbations on Pet1 neuron regulatory element TF binding and accessibility. Our work illustrates the power of DGF for gaining unique insight into how TF missense mutations interfere with developing neuronal GRNs.

Short- and long-term reconfiguration of rat prefrontal cortical networks following single doses of psilocybin
We quantify cellular- and circuit-resolution neural network dynamics following therapeutically relevant doses of the psychedelic psilocybin. Using chronically implanted Neuropixels probes, we recorded local field potentials (LFP) alongside action potentials from hundreds of neurons spanning infralimbic, prelimbic and cingulate subregions of the medial prefrontal cortex of freely-behaving adult rats. Psilocybin (0.3mg/kg or 1mg/kg i.p.) unmasked 100Hz high frequency oscillations that were most pronounced within the infralimbic cortex, persisted for approximately 1h post-injection and were accompanied by decreased net pyramidal cell firing rates and reduced signal complexity. These acute effects were more prominent during resting behaviour than during a sustained attention task. LFP 1-, 2- and 6-days post-psilocybin showed gradually-emerging increases in beta and low-gamma (20-60Hz) power, specific to the infralimbic cortex. These findings reveal features of psychedelic action not readily detectable in human brain imaging, implicating infralimbic network oscillations as potential biomarkers of psychedelic-induced network plasticity over multi-day timescales.

Motor imagery and execution activate similar finger representations that are spatially consistent over time
Finger representations in the sensorimotor cortex can be activated even in the absence of somatosensory input or motor output through mere top-down processes, such as motor imagery. While executed finger movements activate finger representations in the primary sensorimotor cortex that are spatially consistent over time within participants, the stability of top-down activated finger representations remains largely unexplored. Given the increasing use of top-down activated sensorimotor representations to both plan implantation of and control brain-computer interfaces, it is crucial to understand the stability of these representations. Here, we investigated the spatial consistency, and thereby reliability, of finger representations activated through motor imagery in the primary somatosensory and primary motor cortex over time. To assess this, participants performed imagined and executed individual finger movements in two 3T fMRI sessions that were ~2 weeks apart. We observed highly consistent univariate finger-selective activity clusters and multivariate vertex-wise activity patterns within participants over time in both the motor imagery and motor execution task. Using a multivariate across-task decoding approach, we further found that motor execution and motor imagery activate similar finger representations in both the primary somatosensory and primary motor cortex. This demonstrates that motor imagery can be used to identify finger representations related to movement execution. Our findings not only validate the use of top-down processes for brain-computer interface planning and control, but also open up new opportunities for the development of sensorimotor training interventions that do not rely on overt movements.

Classification of Visual Imagery and Imagined Speech EEG based Brain Computer Interfaces using 1D Convolutional Neural Network
Non-invasive brain-computer interfaces (BCI) utilising electroencephalogram (EEG) signals are a current popular, affordable and accessible method for establishing communication paths between the mind and external devices. However, the challenges faced are inter-subject variability, BCI illiteracy and poor machine learning decoding performance. Two emerging intuitive mental paradigms, Visual Imagery (VI) and Imagined Speech (IS) show promise to optimise the development of non-invasive BCIs, which involves the extraction of corresponding neural patterns during the imagined tasks. This study took a comprehensive user-centric approach to build on the current foundation of knowledge on VI and IS EEG-BCIs utilising an adapted 1D-CNN to optimise the classification decoding performance. Twenty healthy participants were assessed for their ability to visualise imagery in their minds and performed the VI and IS mental paradigms in two class conditions "push" and "relax". It was shown that alpha and beta suppression was observed during the "push" condition of VI compared to the "relax" condition, and those that scored higher in the VVIQ had better VI classification accuracy than those who did not. The adapted 1D-CNN model performed well for classification between the two classes "push" and "relax" at 89.3% and 77.87% performance accuracy for VI and IS, respectively. These findings contribute to the current body of work on VI BCI, that it is a dynamic and plausible option compared to standard BCI paradigms, and VI BCI illiteracy could potentially be controlled by VVIQ. It also demonstrated the potential of the 1D-CNN model in classification of VI and IS EEG-BCIs.

A role for fibroblast and mural cell subsets in models of neuropathic pain.
Neuropathic pain is a particularly intractable type of chronic pain that can result from physical nerve damage due to surgery or entrapment. Here, we present data which suggest that a particular subclass of fibroblast and mural cells may be implicated in the sensory neuron dysfunction that is characteristic of this pain state.

In a mouse model of traumatic painful neuropathy, we used RNA sequencing, cell sorting and nerve tissue clearing to study mesenchymal lineage cells. We show that Pdgfrb+ fibroblasts and mural cells are increased in number for at least two months post-nerve damage and express high levels of known and putative pro-algesic mediators, which are further upregulated in neuropathy.

We go on to demonstrate that a human nerve pericyte line releases a selection of these pro-algesic mediators at protein level. Moreover, conditioned media from stimulated human pericytes induces intra-cellular changes in human induced pluripotent stem cell derived sensory neurons; these changes (phosphorylation of the transcription factor STAT3) have been previously linked to sensory neuron activation.

In summary, our data indicate that mesenchymal cell abnormalities should be considered when developing novel strategies to tackle neuropathic pain.

Dopamine regulates the membrane potential and glycine release of AII amacrine cells via D1-like receptor modulation of gap junction coupling.
Dopamine plays a pivotal role in adjusting the flow of information across the retina as luminance changes from night to day. Here we show, under dim photopic conditions, that both dopamine and a D1-like receptor (D1R) agonist hyperpolarized the resting membrane potential (Vm) of AII amacrine cells (AII-ACs). Surprisingly, in the presence of glutamatergic and GABAergic synaptic blockers that isolate glycinergic synapses, D1R agonists are without effect. However, a D1R antagonist depolarized Vm and reduced the input resistance of AII-ACs in wild type mice, but not in Cx36-/- mice. Accordingly, D1R antagonists enhanced tonic glycinergic transmission to type-2 OFF-cone bipolar cells (OFF-CBCs). D1Rs thus adjust the Vm and excitability of AII-ACs and, thereby, the level of glycine release to OFF-CBCs by regulating gap junction coupling with ON cone bipolar cells. Our findings provide insights into how the retina may use dopamine to adapt crossover inhibitory microcircuits during changes in luminance.

M102, a combined NRF2 and HSF-1 activator for neuroprotection in amyotrophic lateral sclerosis.
Abstract: M102 is a central nervous system (CNS) penetrant small molecule electrophile which activates in vivo the NFE2-related factor 2 antioxidant response element (NRF2-ARE) pathway, as well as transcription of heat-shock element (HSE) associated genes. In the TDP-43Q331K transgenic mouse model of ALS dosed subcutaneously at 5mg/kg OD or 2.5mg/kg BD with M102, significant improvements in compound muscle action potential (CMAP) amplitude of hind limb muscles and gait parameters were observed at 6 months of age, with associated target engagement. An oral dose response study of M102 in SOD1G93A transgenic mice showed a dose-dependent improvement in the CMAP of hindlimb muscles which correlated with preservation of lumbar spinal motor neurons at the same time point. These data enabled prediction of human efficacious exposures and doses, which were well within the safety margin predicted from Good Laboratory Practice (GLP) toxicology studies. A parallel program of work in vitro showed that M102 rescued motor neuron survival in co-culture with patient-derived astrocytes from sporadic, C9orf72 and SOD1 cases. Markers of oxidative stress, as well as indices of TDP-43 proteinopathy were also reduced by exposure to M102 in these in vitro models. This comprehensive package of preclinical efficacy data across two mouse models as well as patient-derived astrocyte toxicity assays, provides a strong rationale for clinical evaluation of M102 in ALS patients. Combined with the development of target engagement biomarkers and the completed preclinical toxicology package, a clear translational pathway to testing in ALS patients has been developed.