bioRxiv Subject Collection: Neuroscience's Journal

Changes in Gait Asymmetry May Be Caused by Adaptation of Spinal Reflexes
In a recent human study, we found that adaptive changes in step length asymmetry (SLA) are correlated with similar changes in the H-reflex gains of the leg muscles during split-belt treadmill locomotion. While this observation indicated a closer link between gait asymmetry and spinal reflex adaptation, it did not reveal their causal relationship. To better understand this relationship, here we use a neuromuscular model of human walking whose control relies primarily on spinal reflexes. Subjecting the model to split-belt treadmill locomotion with different combinations of belt speed and reflex gain adaptation patterns, we find that belt speed changes increase the variability in SLA but do not result in consistent SLA patterns as observed in human experiments, whereas reflex gain adaptations do. Furthermore, we find that the model produces SLA patterns similar to healthy adults when its reflex gains are adapted in a way similar to the H-reflex changes we observed in our previous human study. The model also predicts SLA patterns similar to the ones observed for cerebellar degeneration patients when the reflexes do not adapt beyond a sudden dip at the time the ipsilateral belt speed is lowered. Our results suggest that SLA does not arise from imposing belt speed changes but requires the adaptation of the reflex gains, and that the dynamic adaptation of these gains may be an essential part of human gait control when encountering unexpected environment changes such as the uneven speed changes in split-belt treadmill locomotion.

Large-Scale Functional Connectome Fingerprinting for Generalization and Transfer Learning in Neuroimaging
Functional MRI currently supports a limited application space stemming from modest dataset sizes, large interindividual variability and heterogeneity among scanning protocols. These constraints have made it difficult for fMRI researchers to take advantage of modern deep-learning tools that have revolutionized other fields such as NLP, speech transcription, and image recognition. To address these issues, we scaled up functional connectome fingerprinting as a neural network pre-training task, drawing inspiration from speaker recognition research, to learn a generalizable representation of brain function. This approach sets a new high-water mark for neural fingerprinting on a previously unseen scale, across many popular public fMRI datasets (individual recognition over held out scan sessions: 94% on MPI-Leipzig, 94% on NKI-Rockland, 73% on OASIS-3, and 99% on HCP). We show that this representation can also generalize to support accurate neural fingerprinting for completely new datasets and participants not used in training. Finally, we demonstrate that the representation learned by the network encodes features related to individual variability that supports some transfer learning to new tasks. These results open the door for a new generation of clinical applications based on functional imaging data.

The Recurrent Temporal Restricted Boltzmann Machine Captures Neural Assembly Dynamics in Whole-brain Activity
Animal behaviour alternates between stochastic exploration and goal-directed actions, which are generated by the underlying neural dynamics. Previously, we demonstrated that the compositional Restricted Boltzmann Machine (cRBM) can decompose whole-brain activity of larval zebrafish data at the neural level into a small number ($sim$100-200) of assemblies that can account for the stochasticity of the neural activity (van der Plas et al., eLife, 2023). Here we advance this representation by extending to a combined stochastic-dynamical representation to account for both aspects using the Recurrent Temporal RBM (RTRBM) and transfer-learning based on the cRBM estimate. We demonstrate that the functional advantage of the RTRBM is captured in the temporal weights on the hidden units, representing neural assemblies, both in simulated and experimental data. Our results show that the temporal expansion outperforms the stochastic-only cRBM in terms of generalization error and achieves more accurate representation of the moments in time. Lastly, we demonstrate that we can identify the original time-scale of assembly dynamics, by estimating multiple RTRBMs at different temporal resolutions. Together, we propose that RTRBMs are a valuable tool for capturing the combined stochastic and time-predictive dynamics of large-scale data sets.

Investigation of the Relationship Between Calpain and HMGB1/TLR4/NF-KB Signaling Pathway in Multiple Sclerosis and Other Demyelinating Diseases
Background: To develop more effective treatments for demyelinating diseases, it is essential to identify the associated signaling pathways and factors. The objective of this study was to investigate the possible correlation between Calpain-1 (CAPN1) and Calpain-2 (CAPN2) with the HMGB1/TLR4/NF-kappaB signaling pathway and to evaluate the influence of these proteins on Interleukin 17A (IL-17A) and Interleukin 37 (IL-37) cytokines in individuals with recently diagnosed untreated before Multiple Sclerosis (MS) and Neuromyelitis Optica Spectrum Disorder (NMOSD). Methods: In this pilot study, a total of 73 newly diagnosed patients were recruited, including 36 with MS, 9 with NMOSD, and a control group composed of 28 individuals who had Pseudotumour cerebri (unhealthy control). To ensure accuracy and transparency, the demographic and clinical characteristics of all groups were meticulously described. ELISA technique was utilized to compare levels of CAPN1, CAPN2, HMGB1, TLR4, and NF-KB, as well as IL-17A and IL-37 cytokines, between the case groups and the unhealthy control group. The expectation from these findings is to provide valuable insights into the pathophysiological mechanisms of these neurological disorders, possibly opening the door to novel therapeutic perspectives. Results: In patients with MS, the levels of CAPN1 were found to be higher than those in patients with NMOSD and PTS patients. Similarly, the level of CAPN2 was significantly higher in MS patients than in NMOSD patients and higher in PTS patients than in NMOSD patients. There were no notable differences in the levels of HMGB1, TLR4, NF-kappaB, IL-17A, and IL-37 between the groups. Age and gender did not significantly affect any of the parameters. In the MS group, both CAPN1 and CAPN2 showed positive correlation with HMGB1, TLR4, and NF-kappaB levels. Conclusions: It may be suggested that CAPN1 may exhibit greater efficacy than CAPN2 during the initial stages of neuroinflammation. To obtain deeper and more guiding results of the varying levels of CAPN1 and CAPN2, and their relationship with the HMGB1/TLR4/NF-{kappa}B signaling pathway, it is advisable to conduct in-vivo and in-vitro prospective studies featuring CAPN1-specific inhibitors and larger study groups.

Altered nuclear envelope homeostasis is a key pathogenic event in C9ORF72-linked ALS/FTD
ALS and FTD are complex neurodegenerative disorders that primarily affects motor neurons in the brain and spinal cord, and cortical neurons in the frontal lobe. Although the pathogenesis of ALS/FTD is unclear, recent research spotlights nucleocytoplasmic transport impairment, DNA damage, and nuclear abnormalities as drivers of neuronal death. In this study, we show that loss of nuclear envelope (NE) integrity is a key pathology associated with nuclear pore complex (NPC) injury in C9ORF72 mutant neurons. Importantly, we show that mechanical stresses generated by cytoskeletal forces on the NE can lead to NPC injury, loss of nuclear integrity, and accumulation of DNA damage. Importantly, we demonstrate that restoring NE tensional homeostasis, by disconnecting the nucleus from the cytoskeleton, can rescue NPC injury and reduce DNA damage in C9ORF72 mutant cells. Together, our data suggest that modulation of NE homeostasis and repair may represent a novel and promising therapeutic target for ALS/FTD.

Disrupted mitochondrial response to nutrients is a presymptomatic event in the cortex of the APPSAA knock-in mouse model of Alzheimer disease
Introduction: Reduced brain energy metabolism, mTOR dysregulation, and extracellular amyloid-{beta} oligomer (xcA{beta}O) buildup characterize AD; how they collectively promote neurodegeneration is poorly understood. We previously reported that xcA{beta}Os inhibit Nutrient-induced Mitochondrial Activity (NiMA) in cultured neurons. We now report NiMA disruption in vivo. Methods: Brain energy metabolism and oxygen consumption were recorded in APPSAA/+ mice using two-photon fluorescence lifetime imaging and multiparametric photoacoustic microscopy. Results: NiMA is inhibited in APPSAA/+ mice before other defects are detected in these amyloid-{beta}-producing animals that do not overexpress APP or contain foreign DNA inserts into genomic DNA. GSK3{beta} signals through mTORC1 to regulate NiMA independently of mitochondrial biogenesis. Inhibition of GSK3{beta} with lithium or TWS119 stimulates NiMA in cultured human neurons, and mitochondrial activity and oxygen consumption in APPSAA mice. Conclusion: NiMA disruption in vivo occurs before histopathological changes and cognitive decline in APPSAA mice, and may represent an early stage in human AD.

Automated neuropil segmentation of fluorescent images for Drosophila brains
The brain atlas, which provides information about the distribution of genes, proteins, neurons, or anatomical regions in the brain, plays a crucial role in contemporary neuroscience research. To analyze the spatial distribution of those substances based on images from different brain samples, we often need to warp and register individual brain images to a standard brain template. However, the process of warping and registration often leads to spatial errors, thereby severely reducing the accuracy of the analysis. To address this issue, we develop an automated method for segmenting neuropils in the Drosophila brain using fluorescence images from the FlyCircuit database. This technique allows future brain atlas studies to be conducted accurately at the individual level without warping and aligning to a standard brain template. Our method, LYNSU (Locating by YOLO and Segmenting by U-Net), consists of two stages. In the first stage, we use the YOLOv7 model to quickly locate neuropils and rapidly extract small-scale 3D images as input for the second stage model. This stage achieves a 99.4% accuracy rate in neuropil localization. In the second stage, we employ the 3D U-Net model to segment neuropils. LYNSU can achieve high accuracy in segmentation using a small training set consisting of images from merely 16 brains. We demonstrate LYNSU on six distinct neuropils or structure, achieving a high segmentation accuracy, which was comparable to professional manual annotations with a 3D Intersection-over-Union(IoU) reaching up to 0.869. Most notably, our method takes only about 7 seconds to segment a neuropil while achieving a similar level of performance as the human annotators. The results indicate the potential of the proposed method in high-throughput connectomics construction for Drosophila brain optical imaging.

Excitation/Inhibition balance relates to cognitive function and gene expression in Temporal Lobe Epilepsy: an hdEEG assessment with aperiodic exponent
Patients with epilepsy are characterized by a dysregulation of excitation-inhibition balance (E/I). The assessment of E/I may inform clinicians during the diagnosis and therapy management, even though it is rarely performed. An accessible measure of the E/I of the brain represents a clinically relevant feature. Here we exploited the exponent of the aperiodic component of the power spectrum of EEG signal as a noninvasive and cost-effective proxy of the E/I balance. We recorded resting-state activity with high-density EEG from 65 patients with temporal lobe epilepsy (TLE) and 35 controls. We extracted the exponent of the aperiodic fit of the power spectrum from source-reconstructed EEG and tested differences between TLE and controls. Spearman correlation was performed between the exponent and clinical variables (age of onset, epilepsy duration and neuropsychology) and cortical expression of epilepsy-related genes derived from Human Allen Brain Atlas. Patients with TLE showed a significantly larger exponent, corresponding to an inhibition directed E/I balance, in bilateral frontal and temporal regions. Lower E/I in the left entorhinal, and bilateral dorsolateral prefrontal cortices corresponded to a lower performance of short term verbal memory. Limited to TLE, we detected a significant correlation between the exponent and the cortical expression of GABRA1, GRIN2A, GABRD, GABRG2, KCNA2and PDYN. EEG aperiodic exponent maps the E/I balance non-invasively in patients with epilepsy and reveals a tight relationship between altered E/I patterns, cognition and genetics.

Neonatal Amygdala Mean Diffusivity: A Potential Predictor of Emotional Face Perception
The ability to differentiate between different facial expressions is an important part of human social and emotional development that begins in infancy. Studies have shown that within the first year of life, infants develop a distinctive attentional bias towards fearful facial expressions. Investigations into the neural basis for this bias have highlighted the significance of the amygdala. The amygdala's role in directing attention towards fearful facial expressions underscores its importance in early emotional development, significantly influencing how infants interpret and react to facial expressions. To date, no studies have been conducted to investigate the associations between the amygdala microstructure and infants' perception of emotional faces. This study aimed to elucidate this relationship while also investigating whether this association is sex specific. We measured the amygdala microstructural properties using diffusion tensor imaging mean diffusivity (MD) measurements in 40 healthy infants aged 2 to 5 weeks. Eye tracking was used to assess attention disengagement from fearful vs. non-fearful (happy and neutral) facial expressions as well as scrambled non-face control picture at 8 months. Generally, infants were age-typically less likely to disengage from fearful faces than from non-fearful faces towards salient distractors. A significant negative association was observed between the right amygdala MD measures and disengagement probability from fearful faces in the overall sample. Moreover, there was a positive association between the bilateral amygdala MD measures and the disengagement probability from scrambled non-face control picture in girls. These results indicate that the amygdala MD is associated with attention disengagement processes already in infancy, both in fear processing and in non-emotional conditions. Specifically, these findings highlight the role of the amygdala microstructure in modulating attentional processes, which may have implications for emotional regulation and susceptibility to emotional dysregulation later in life.

Multiple Loci for Foveolar Vision in Macaque Monkey
A common tenet of neural sensory representation is that species-specific behaviors are reflected in specialized brain organizations. In humans and nonhuman primates, the central one degree of vision is processed by the foveola, a retinal structure which comprises a high density of photoreceptors and is crucial for primate-specific high acuity vision, color vision, and gaze-directed visual attention. In this study, we have developed high spatial resolution ultrahigh field 7T fMRI methods for functional mapping of foveolar visual cortex in awake monkeys. We provide evidence that, in the ventral pathway (V1-V4 and TEO), viewing of a central small spot elicits a ring of multiple (at least 8) foveolar representations per hemisphere. This ring surrounds a large area called the "foveolar core". This is an area populated by millimeter-scale functional domains sensitive to fine stimuli and high spatial frequencies, consistent with foveolar visual acuity, as well as color and achromatic information, and motion. The unique position of the foveolar core suggests it may be a hub subserving higher order needs of foveolar function, such as integrating different spatial scales of representation, integrating local and global features in object perception, and bringing together the four quadrants of visual space. Thus, this elaborate re-representation of central vision signifies a cortical specialization for various foveation behaviors.

Prefrontal cortex temporally multiplexes slow and fast dynamics in value learning and memory
Previous studies have revealed segregated circuitries in basal ganglia for fast learning that enables value adaptability and slow forgetting which underlies stable value memories. However, the mechanisms mediating the conflict between value adaptability vs stability remain unknown. Using a reinforcement learning paradigm involving a brief value reversal for objects with previously stable values, we predicted and confirmed a novel behavioral manifestation of the conflict between adaptability vs stability namely the spontaneous recovery of old values in macaque monkeys. Furthermore, we found that individual neurons in ventrolateral prefrontal cortex (vlPFC) temporally multiplexed slow and fast processes in their early and late responses to objects. The local field potential in vlPFC also reflected the two-rate system. These findings implicate vlPFC as a plexus for the interactions between adaptability vs stability in reinforcement learning and suggest spontaneous recovery of past values caused by a two-rate system to mediate relapse to old habits.

Limb, not touch location, is coded in 3D space
Skin location of touch is said to be recoded, by default, into a 3D-spatial location. Here, human participants received tactile stimulus pairs on a common or on two different limbs and judged either whether stimuli had occurred on the same limb, or on a common side of space. Misaligning skin and 3D-spatial codes through limb crossing had a stronger effect on spatial than on limb judgments, contradicting the notion that the 3D location of touch is readily available. Additionally, crossing effects were significantly reduced when stimulus pairs lay in a common dermatome, suggesting that limb choice was based on anatomical, not 3D-spatial, stimulus information. These results suggest that it is limbs, not touch, which are coded in 3D space. Commonly observed errors in reporting where touch occurred on the body depend on confusion due to where the touched limb usually resides with the spatial layout of the currently action-relevant limbs.

Bundle Analytics based Data Harmonization for Multi-Site Diffusion MRI Tractometry
The neural pathways of the living human brain can be tracked using diffusion MRI-based tractometry. Along-tract statistical analysis of microstructural metrics can reveal the effects of neurological and psychiatric diseases with 3D spatial precision. To maximize statistical power to detect disease effects and factors that influence them, data from multiple sites and scanners must often be combined, yet scanning protocols and hardware may vary widely. For simple scalar metrics, data harmonization methods - such as ComBat and its variants - allow modeling of disease effects on derived brain metrics, while adjusting for effects of scanning site or protocol. Here we extend this method to pointwise analyses of 3D fiber bundles, by integrating ComBat into the BUndle ANalytics (BUAN) tractometry pipeline. In a study of the effects of mild cognitive impairment (MCI) and Alzheimer's disease (AD) on 38 white matter tracts, we merge data from 7 different scanning protocols used in the Alzheimer's Disease Neuroimaging Initiative, which vary in voxel size and angular resolution. By incorporating ComBat harmonization, we model site- and scanner-specific effects, ensuring the reliability and comparability of results by mitigating confounding variables. We also evaluate choices that arise in extending batch adjustment to tracts, such as the regions used to estimate the correction. We also compare the approach to the simpler approach of modeling the site as a random effect. To the best of our knowledge, this is one of the first applications to adapt harmonization to 3D tractometry.

Visual stimulation drives retinotopic acetylcholine release in the mouse visual cortex
Cholinergic signaling is involved with a variety of brain functions including learning and memory, attention, and behavioral state modulation. The spatiotemporal characteristics of neocortical acetylcholine (ACh) release in response to sensory inputs are poorly understood, but a lack of intra-region topographic organization of cholinergic projections from the basal forebrain has suggested diffuse release patterns and volume transmission. Here, we use mesoscopic imaging of fluorescent ACh sensors to show that visual stimulation results in ACh release patterns that conform to a retinotopic map of visual space in the mouse primary visual cortex, suggesting new modes of functional cholinergic signaling in cortical circuits.

Deriving the cone fundamentals: a subspace intersection method
Two ideas, proposed by Thomas Young and James Clerk Maxwell, form the foundations of color science: (1) Three types of retinal receptors encode light under daytime conditions, and (2) color matching experiments establish the critical spectral properties of this encoding. Experimental quantification of these ideas are used in international color standards. But, for many years the field did not reach consensus on the spectral properties of the biological substrate of color matching: the sensitivity of the in situ cones (cone fundamentals). By combining auxiliary data (thresholds, inert pigment analyses), complex calculations, and color matching from genetically analyzed dichromats, the human cone fundamentals have now been standardized. Here we describe a new computational method to estimate the cone fundamentals using only color matching from dichromatic observers. We show that it is not necessary to include data from trichromatic observers in the analysis or to know the primary lights used in the matching experiments. Remarkably, it is even possible to estimate the fundamentals by combining data from experiments using different, unknown primaries. We then suggest how the new method may be applied to color management in modern image systems.

Acquiring social safety engages oxytocin neurons in the supraoptic nucleus - role of Magel2 deficiency
Introduction: Exposure to social trauma may alter engagement with both fear-related and unrelated social stimuli long after. Intriguingly, how simultaneous discrimination of social fear and safety is affected in neurodevelopmental conditions like autism remains underexplored. The role of the neuropeptide oxytocin is established in social behaviors, and yet unexplored during such a challenge post-social trauma. Methods: Using Magel2-knockout mice, an animal model of Prader Willi Syndrome (PWS) and autism spectrum disorders, we tested memory of social fear and safety after a modified social fear conditioning task. Additionally, we tracked the activity of oxytocin neurons in the supraoptic (SON) and paraventricular (PVN) nuclei of the hypothalamus by fibre photometry, as animals were simultaneously presented with a choice between a fear and safe social cue during recall. Results: Male Magel2 KO mice trained to fear females with electrical footshocks avoided both unfamiliar females and males during recalls, lasting even a week post-conditioning. On the contrary, trained Magel2 WT avoided only females during recalls, lasting days rather than a week postconditioning. Inability to overcome social fear and avoidance of social safety in Magel2 KO mice were associated with reduced engagement of oxytocin neurons in the SON, but not the PVN. Conclusion: In a preclinical model of PWS, we demonstrated region-specific deficit in oxytocin activity associated with behavioral generalization of social fear to social safety. Insights from this study add to our understanding of oxytocin action in the brain at the intersection of social trauma, PWS and related autism spectrum disorders.