Biology subjects

Sur, M.

Publications and source records attributed to Sur, M..

At least 19 recordsLinked to original sources

bioRxiv2026

Eye-specific drive and binocular suppression in mouse visual cortex during critical period development

A key function of cortical sensory circuits is to integrate information from multiple sources to build a unified representation of the external environment. The binocular region of the mouse primary visual cortex (bV1) is a valuable model for studying sensory integration, as visual response properties of inputs from the contralateral (contra) and ipsilateral (ipsi) eye onto bV1 neurons become matched to one another over development. While the alignment of visual responses from each eye during this developmental critical period has been well-characterized, it remains unclear how this interocular alignment influences, and is influenced by, the binocularly driven responses of bV1 neurons. Here, we recorded the monocular and binocular visual response properties of layer 2/3 bV1 neurons by performing chronic two-photon calcium imaging at multiple time points spanning the ocular-dominance critical period. We found that ipsi driven responses strengthened in bV1 neurons alongside a concurrent increase in binocular suppression. Binocular tuning properties were more strongly aligned with contra responses than with ipsi responses throughout development but became progressively better matched to ipsi responses. Additionally, the explained variance of ipsi input to predicting binocular tuning increased. In chronically tracked bV1 neurons, the preferred orientation of ipsi eye driven responses was less stable than that of contra eye driven responses, and in well-tuned stable binocular neurons, ipsi and contra preferences became progressively better matched. Using population-based decoding analyses, we found that while monocular and binocular visual encoding remained stable, there was increased generalizability between visual encoding by binocular and ipsi eye driven responses. Overall, our data suggest a reciprocal interaction between ipsi development and binocular responses, such that ipsi visual inputs dynamically structure excitatory as well as putative inhibitory drive onto binocular neurons during the critical period to generate the circuitry that implements integrated binocular vision.

neuroscience↗

bioRxiv2026

Internal state dynamically gates task-specific attractor dynamics in prefrontal cortex

Internal states such as motivation and task engagement influence cognitive functions. Working memory, which maintains information over time, is an essential component of cognition and is modulated by motivation. Here, we show motivational states modulated attractor dynamics that supported working memory. Combining population recordings from mouse medial prefrontal cortex (mPFC) with data-constrained recurrent neural network (RNN) modeling, we found task engagement selectively modulated attractor dynamics within a memory-maintenance subspace, while stimulus-evoked responses remained intact. Reverse-engineering the RNNs revealed that task engagement reorganized the dynamical landscape by stabilizing memory-specific attractors. Specifically, task engagement modulated interactions between neurons to change the attractor dynamics. Finally, gradual changes in behavioral engagement were predicted by continuous modulation of attractor geometry in RNNs and mPFC. Together, these results suggest that internal state modulate working memory function by controlling the dynamical regime of mPFC circuits, providing a mechanistic link between internal state, neural dynamics, and cognitive function.

neuroscience↗

bioRxiv2026

Decomposing the modulation of interactions between neuronal populations

Identifying subpopulations of neurons that interact with each other from simultaneous recordings of populations of many neurons is key for understanding across-brain communication with cellular resolution. Recent work identified communication subspaces, which capture additive interactions between pairs of high-dimensional neural populations through a small number of source and target activity patterns. However, no current method captures how a third, potentially multivariate variable - such as behavioral state or the activity of a third population - modulates these interactions. Here we extend the communication subspace framework by parameterizing modulation as a low-rank tensor. This identifies multiplicative interaction channels (MICs), defined as triplets of source, target, and modulator activity patterns, in which the modulator pattern gates the source-target interaction. We derive MICs as a bilinear perturbation of reduced-rank regression. We develop a hierarchical fitting pipeline and provide a closed-form decomposition that quantifies whether modulation reshapes the modulator-averaged baseline interaction, recruits private dimensions of one population, or opens new interactions. In simulations, MICs reliably recover the presence and geometry of ground-truth modulation even in the high-dimensional, low-sample regime. Applying MICs to simultaneous calcium imaging of prefrontal axons and interneurons in the visual cortex revealed that behavioral state asymmetrically modulates top-down interactions, reconfiguring the patterns of prefrontal projections that interact with a stable set of visual interneuron activity patterns. By providing an efficient and compact characterization of modulatory interactions, MICs enable asking new questions about how potentially high-dimensional variables shape interactions between neural populations.

neuroscience↗

bioRxiv2026

CHARIOT-AAV: Conjugation of diverse vectors to adeno-associated viruses for delivery of large genes

Systemic, tissue-specific delivery of large transgenes exceeding the packaging capacity of adeno-associated viruses (AAVs) remains a key translational challenge for molecular therapeutics. Vectors with larger capacities, such as lentiviral vectors (LVVs) and lipid nanoparticles (LNPs), often lack adjustable, tissue-specific tropisms. Here we report CHARIOT-AAV (Crosslinked Hybrid Architectures for Robust, Interchangeable, and Organ-specific Targeting with AAV), a platform where diverse delivery vectors are conjugated to AAVs, thereby achieving tissue-specific tropism of AAVs and expanded cargo capacity. AAV-AAV conjugates packaging split SpCas9 constructs in AAV.CAP-B10 capsids demonstrate a [~]2-fold increase in brain gene editing efficiency over unconjugated AAV cocktails after intravenous injection. In addition to AAV-AAV conjugates, AAV-LVV and AAV-LNP conjugates achieve AAV-guided delivery of genetic payloads to target cells. Furthermore, AAV-LNP conjugates enable systemic delivery of mRNAs to brain endothelial cells. CHARIOT-AAV thus provides a modular platform for systemic, tissue-specific delivery of diverse therapeutics beyond the limits of individual vectors.

bioengineering↗

bioRxiv2026

IGF1 peptide targets Rett Syndrome astrocytes to degrade IGF binding protein, rescue synaptogenesis and restore mitochondrial function

Rett syndrome (RTT), a severe neurodevelopmental disorder caused by mutations in MECP2, leads to profound synaptic and circuit deficits in the brain. While neurons have historically been the focus of RTT pathology, emerging evidence implicates astrocytes in non-cell autonomous mechanisms that impair synaptic structure, function and development. Here, we uncover a central role for astrocyte-secreted IGFBP2 in mediating these deficits and demonstrate that treatment with an IGF1-derived peptide restores synapse formation by promoting IGFBP2 degradation. Using an indirect astrocyte-neuron co-culture system, we show that astrocytes derived from RTT model mice suppress excitatory synapse formation in wild-type neurons and that this impairment is reversed when RTT astrocytes are treated with IGF1(1-3) peptide. Proteomic analysis reveals elevated levels of IGFBP2 in RTT astrocytes and their conditioned media. IGF1(1-3) peptide treatment leads to proteasomal degradation of IGFBP2, increasing IGF1 bioavailability, restoring mitochondrial function, and enhancing downstream PI3K/Akt signaling in neurons. Our data define a molecular mechanism by which astrocyte dysfunction in RTT can be rescued and provide a mechanistic basis for the therapeutic efficacy of IGF1(1-3) peptide, including Trofinetide, an FDA-approved IGF1 peptide mimetic, in RTT. Significance StatementAstrocyte dysfunction is increasingly recognized as a contributor to neurodevelopmental disorders, yet their precise mechanisms of action remain elusive. Here, we identify IGFBP2 as a key astrocyte-derived inhibitor of synaptogenesis in Rett syndrome. We show that an IGF1-derived peptide, IGF1(1-3), depletes IGFBP2 via proteasomal degradation. This restores IGF1 bioavailability and rescues synaptic function in a non-cell-autonomous manner. These findings provide a mechanistic explanation for the clinical efficacy of IGF1 peptide and its mimetics in Rett syndrome, and highlight astrocytes as rational therapeutic targets in neurodevelopmental and other disorders.

neuroscience↗

bioRxiv2025

Reusable modular architecture enables flexible cognitive operations

Complex behaviors are thought to be built by combining simpler cognitive components. Computational modeling has shown that artificial neural networks can perform a variety of tasks by flexibly combining functional modules, each specialized for a specific computation, to construct a complex task. However, it is unknown whether reusable modular networks are found in the brain. Here, we show that mice performing a delayed match-to-sample with delayed report (DMS-dr) task reuse neuronal subspaces that were specialized for stimulus processing and memory maintenance. These subspaces were reused during the task to represent new stimulus inputs and different types of memories, respectively. Clustering analyses showed each subspace was supported by a distinct cluster of neurons in prefrontal cortex and parietal cortex. Studying artificial recurrent networks constrained to neural data found silencing specific clusters disrupted specific computations, consistent with a modular and reusable organization. Altogether, our findings show that the brain can flexibly reuse computational components to perform a complex cognitive task.

animal behavior and cognition↗

bioRxiv2025

Assessment of the adverse effects of pollution on farmland bird diversity in contemporary agricultural landscapes

The escalating global demand for food has intensified agricultural practices, leading to substantial changes in land use. This transformation poses a threat to farmland biodiversity, compounded by the presence of pollutants from anthropogenic activities. While the impact of specific pollutants is known in controlled environments, their compounded effects under field conditions remain largely unexplored. We investigated correlations between farmland bird distribution, landscape features, land use patterns, and anthropogenic pressures, including nutrient pollution, acidifying compounds, and synthetic chemicals. Using distribution maps of the Netherlands at a 1x1 km{superscript 2} grid cell, we analyzed the association of farmland bird species richness and abundance with landscape characteristics and varying levels of exposures to unintended pollutants. We also compared species richness distribution patterns between 1998 and 2018. We found a strong negative relationship between farmland bird species richness and abundance with atmospheric deposition of inorganic nitrogen (NHx, NOy). Furthermore, mixed associations were observed between farmland birds and local toxic pressure variation in surface waters, with consistent relationships to industrial chemicals (negative) and products of combustion (positive). Lastly, change in species richness from 1998 and 2018, showed that many of the relationships observed now were already evident two decades ago, with recent declines in species richness concentrated in landscapes hosting a considerable number of species, and low nitrogen deposition grid cells. We conclude that although it is likely that there is some direct negative effects of pollutants on farmland birds, it is reasonable to also assume that the identified relationships are proxies for the overarching intensity of farming, human disturbance, and broader landscape changes. Our study highlights a) a possible role of synthetic pollutants and acidifying eutrophicating substances in farmland bird decline b) the need for well-designed field studies to complement correlative evidence from big data approaches such as ours to enhance our understanding and c) the broader implications for sustainable land management, emphasizing the importance of a holistic approach in addressing the intricate relationships between pollutants and landscape changes.

ecology↗

bioRxiv2025

Persistent Disruptions in Prefrontal Connectivity Despite Behavioral Rescue by Environmental Enrichment in a Mouse Model of Rett Syndrome

Rett Syndrome, a neurodevelopmental disorder caused by loss-of-function mutations in the MECP2 gene, is characterized by severe motor, cognitive and emotional impairments. Some of the deficits may result from changes in cortical connections, especially downstream projections of the prefrontal cortex, which may also be targets of restoration following rearing conditions such as environmental enrichment that alleviate specific symptoms. Here, using a heterozygous Mecp2+/- female mouse model closely analogous to human Rett Syndrome, we investigated the impact of early environmental enrichment on behavioral deficits and prefrontal cortex connectivity. Behavioral analyses revealed that enriched housing rescued fine motor deficits and reduced anxiety, with enrichment-housed Mecp2+/- mice performing comparably to wild-type (WT) controls in rotarod and open field assays. Anatomical mapping of top-down anterior cingulate cortex (ACA) projections demonstrated altered prefrontal cortex connectivity in Mecp2+/- mice, with increased axonal density in the somatosensory cortex and decreased density in the motor cortex compared to WT controls. ACA axons revealed shifts in hemispheric distribution, particularly in the medial network regions, with Mecp2+/-mice exhibiting reduced ipsilateral dominance. These changes were unaffected by enriched housing, suggesting that structural abnormalities in prefrontal cortex connectivity persist despite behavioral improvements. Enriched housing rescued brain-derived neurotrophic factor (BDNF) levels in the hippocampus but failed to restore BDNF levels in the prefrontal cortex, consistent with the persistent deficits observed in prefrontal axonal projections. These findings highlight the focal nature of changes induced by reduction of MeCP2 and by exposure to environmental enrichment, and suggest that environmental enrichment starting in adolescence can alleviate behavioral deficits without reversing abnormalities in large-scale cortical connectivity.

neuroscience↗

bioRxiv2024

Sensory History Drives Adaptive Neural Geometry in LP/Pulvinar-Prefrontal Cortex Circuits

Prior expectations guide attention and support perceptual filtering for efficient processing during decision-making. Here we show that during a visual discrimination task, mice use prior stimulus history to guide ongoing choices by estimating differences in evidence between consecutive trials (|{Delta}Dir|). The thalamic lateral posterior (LP)/pulvinar nucleus provides robust inputs to the Anterior Cingulate Cortex (ACC), which has been implicated in selective attention and predictive processing, but the function of the LP-ACC projection is unknown. We found that optogenetic manipulations of LP-ACC axons disrupted animals ability to effectively estimate and use information across stimulus history, leading to |{Delta}Dir|-dependent biases. Two-photon calcium imaging of LP-ACC axons revealed an engagement-dependent low-dimensional organization of stimuli along a curved manifold. This representation was scaled by |{Delta}Dir| in a manner that emphasized greater deviations from prior evidence. Thus, our work identifies the LP-ACC pathway as essential for selecting and evaluating stimuli relative to prior evidence to guide decisions.

neuroscience↗

bioRxiv2024

Astrocytic modulation of population encoding in mouse visual cortex via GABA transporter 3 revealed by multiplexed CRISPR/Cas9 gene editing

Astrocytes, which are increasingly recognized as pivotal constituents of brain circuits governing a wide range of functions, express GABA transporter 3 (Gat3), an astrocyte-specific GABA transporter responsible for maintenance of extra-synaptic GABA levels. Here, we examined the functional role of Gat3 in astrocyte-mediated modulation of neuronal activity and information encoding. First, we developed a multiplexed CRISPR construct applicable for effective genetic ablation of Gat3 in the visual cortex of adult mice. Using in vivo two-photon calcium imaging of visual cortex neurons in Gat3 knockout mice, we observed changes in spontaneous and visually driven single neuronal response properties such as response magnitudes and trial-to-trial variability. Gat3 knockout exerted a pronounced influence on population-level neuronal activity, altering the response dynamics of neuronal populations and impairing their ability to accurately represent stimulus information. These findings demonstrate that Gat3 in astrocytes profoundly shapes the sensory information encoding capacity of neurons and networks within the visual cortex.

neuroscience↗

bioRxiv2024

Cortical norepinephrine-astrocyte signaling critically mediates learned behavior

Updating behavior based on feedback from the environment is a crucial means by which organisms learn and develop optimal behavioral strategies1-3. Norepinephrine (NE) release from the locus coeruleus (LC) has been shown to mediate learned behaviors4-6 such that in a task with graded stimulus uncertainty and performance, a high level of NE released after an unexpected outcome causes adaptations in subsequent behavior7. Yet, how the transient activity of LC-NE neurons, lasting tens of milliseconds, alters neuronal activity and influences behavior several seconds later is unclear. Here, we show that NE released after an unexpected outcome acts directly on cortical astrocytes via 1 adrenergic (Adra1a) receptors to elicit sustained increases in intracellular calcium. Chemogenetic blockade of astrocytic calcium dynamics prevents trial-to-trial behavioral adaptation. NE stimulation of astrocytes elicits ATP release, and imaging ATP levels in the cortex reveals an increase in extracellular ATP in response to an unexpected outcome. Blocking ATP-driven signaling to neuronal adenosine A1 receptors also prevents post-reinforcement behavioral adaptation. Finally, high density neuronal recordings in prefrontal cortex reveal that a surprising outcome alters the neuronal representation of the stimulus on the subsequent trial without sustained changes in cortical activity; blocking either astrocyte calcium dynamics or A1 receptors occludes these post-reinforcement changes in single-neuron and population neuronal encoding of task variables underlying behavioral changes. Together, these data demonstrate that astrocytes play an essential role in norepinephrine-driven learned behavior: they have prolonged calcium responses to transient norepinephrine release and convey task-relevant reinforcement information across behavioral intervals, enabling selective updating of neuronal task representations to support adaptive behavior.

neuroscience↗

bioRxiv2024

miR126-mediated impaired vascular integrity in Rett syndrome

Rett syndrome (RTT) is a neurodevelopmental disorder that is caused by mutations in melty-CpG binding protein 2 (MeCP2). MeCP2 is a non-cell type-specific DNA binding protein, and its mutation influences not only neural cells but also non-neural cells in the brain, including vasculature associated with endothelial cells. Vascular integrity is crucial for maintaining brain homeostasis, and its alteration may be linked to the pathology of neurodegenerative disease, but a non-neurogenic effect, especially the relationship between vascular alternation and Rett syndrome pathogenesis, has not been shown. Here, we recapitulate a microvascular network using Rett syndrome patient-derived induced pluripotent stem (iPS) cells that carry MeCP2[R306C] mutation to investigate early developmental vascular impact. To expedite endothelial cell differentiation, doxycycline (DOX)-inducible ETV2 expression vectors were inserted into the AAVS1 locus of Rett syndrome patient-derived iPS cells and its isogenic control by CRISPR/Cas9. With these endothelial cells, we established a disease microvascular network (Rett-dMVNs) and observed higher permeability in the Rett-dMVNs compared to isogenic controls, indicating altered barrier function by MeCP2 mutation. Furthermore, we unveiled that hyperpermeability is involved in the upregulation of miR126-3p in Rett syndrome patient-derived endothelial cells by microRNA profiling and RNAseq, and rescue of miR126-3p level can recover their phenotype. We discover miR126-3p-mediated vascular impairment in Rett syndrome patients and suggest the potential application of these findings for translational medicine.

cell biology↗

bioRxiv2024

In Vivo Optical Clearing of Mammalian Brain

Established methods for imaging the living mammalian brain have, to date, taken the brains optical properties as fixed; we here demonstrate that it is possible to modify the optical properties of the brain itself to significantly enhance at-depth imaging while preserving native physiology. Using a small amount of any of several biocompatible materials to raise the refractive index of solutions superfusing the brain prior to imaging, we could increase several-fold the signals from the deepest cells normally visible and, under both one-photon and two-photon imaging, visualize cells previously too dim to see. The enhancement was observed for both anatomical and functional fluorescent reporters across a broad range of emission wavelengths. Importantly, visual tuning properties of cortical neurons in awake mice, and electrophysiological properties of neurons assessed ex vivo, were not altered by this procedure.

bioengineering↗

bioRxiv2024

Early differential impact of MeCP2 mutations on functional networks in Rett syndrome patient-derived human cerebral organoids

Human cerebral organoids derived from induced pluripotent stem cells can recapture early developmental processes and reveal changes involving neurodevelopmental disorders. Mutations in the X-linked methyl-CpG binding protein 2 (MECP2) gene are associated with Rett syndrome, and disease severity varies depending on the location and type of mutation. Here, we focused on neuronal activity in Rett syndrome patient-derived organoids, analyzing two types of MeCP2 mutations - a missense mutation (R306C) and a truncating mutation (V247X) - using calcium imaging with three-photon microscopy. Compared to isogenic controls, we found abnormal neuronal activity in Rett organoids and altered network function based on graph theoretic analyses, with V247X mutations impacting functional responses and connectivity more severely than R306C mutations. These changes paralleled EEG data obtained from patients with comparable mutations. Labeling DLX promoter-driven inhibitory neurons demonstrated differences in activity and functional connectivity of inhibitory and excitatory neurons in the two types of mutation. Transcriptomic analyses revealed HDAC2-associated impairment in R306C organoids and decreased GABAA receptor expression in excitatory neurons in V247X organoids. These findings demonstrate mutation-specific mechanisms of vulnerability in Rett syndrome and suggest targeted strategies for their treatment.

bioengineering↗

bioRxiv2024

Prefrontal Cortex subregions provide distinct visual and behavioral feedback modulation to the Primary Visual Cortex

The mammalian Prefrontal Cortex (PFC) has been suggested to modulate sensory information processing across multiple cortical regions via long-range axonal projections. These axonal projections arise from PFC subregions with unique brain-wide connectivity and functional repertoires, which may provide the architecture for modular feedback intended to shape sensory processing. Here, we used axonal tracing, axonal and somatic 2-photon calcium imaging, and chemogenetic manipulations in mice to delineate how projections from the Anterior Cingulate Cortex (ACA) and ventrolateral Orbitofrontal Cortex (ORB) of the PFC modulate sensory processing in the primary Visual Cortex (VISp) across behavioral states. Structurally, we found that ACA and ORB had distinct patterning of projections across cortical regions and layers, monosynaptically targeting both excitatory and inhibitory neuronal cell types in these areas. ACA axons in VISp had a stronger representation of visual stimulus information than ORB axons, but both projections showed non-visual, behavior-dependent activity. ACA input to VISp enhanced the encoding of visual stimuli by VISp neurons, and modulation of visual responses scaled with arousal. On the other hand, ORB input shaped movement and arousal related modulation of VISp visual responses, but surprisingly reduced the encoding of high-contrast visual stimuli. Thus, ACA and ORB feedback have separable projection patterns and encode distinct visual and behavioral information, putatively providing the substrate for their unique effects on visual representations and behavioral modulation in VISp. Our results offer a refined model of cortical hierarchy and its impact on sensory information processing, whereby specific PFC projections contribute uniquely to VISp activity during discrete behavioral states.

neuroscience↗

bioRxiv2024

The noncoding circular RNA circHomer1 regulates developmental experience-dependent plasticity in mouse visual cortex

Circular RNAs (circRNAs) are a class of closed-loop, single stranded RNAs whose expression is particularly enriched in the brain. Despite this enrichment and evidence that the expression of circRNAs are altered by synaptic development and in response to synaptic plasticity in vitro, the regulation by and function of the majority of circRNAs in experience-dependent plasticity in vivo remain unexplored. Here, we employed transcriptome-wide analysis comparing differential expression of both mRNAs and circRNAs in juvenile mouse primary visual cortex (V1) following monocular deprivation (MD), a model of experience-dependent developmental plasticity. Among the differentially expressed mRNAs and circRNAs following 3-day MD, the circular and the activity-dependent mRNA forms of the Homer1 gene, circHomer1 and Homer1a respectively, were of interest as their expression changed in opposite directions: circHomer1 expression increased while the expression of Homer1a decreased following 3-day MD. Knockdown of circHomer1 delayed the depression of closed-eye responses normally observed after 3-day MD. circHomer1-knockdown also led to a reduction in average dendritic spine size prior to MD but critically there was no further reduction after 3-day MD, consistent with impaired structural plasticity. circHomer1-knockdown also prevented the reduction of surface AMPA receptors after 3-day MD. Synapse-localized puncta of the AMPA receptor endocytic protein Arc increased in volume after MD but were smaller in circHomer1-knockdown neurons, suggesting that circHomer1 knockdown impairs experience-dependent AMPA receptor endocytosis. Thus, the expression of multiple circRNAs are regulated by experience-dependent developmental plasticity, and our findings highlight the essential role of circHomer1 in V1 synaptic development and experience-dependent plasticity. Significance StatementCircular RNAs (circRNAs) are a class of closed-loop RNAs formed through back-splicing of exon and/or intron junctions. Initially considered as byproducts of aberrant RNA splicing with limited function, recent studies have implicated circRNAs in various neurological disorders. Despite their abundant expression in the brain, the role of circRNAs in experience-dependent plasticity remains poorly understood. We conducted an in vivo transcriptome analysis of circRNAs whose expression was regulated by experience-dependent plasticity in the mouse visual cortex and highlight circHomer1, a circRNA derived from the Homer1 gene, as a circRNA critical for synaptic development and experience-dependent plasticity in vivo.

neuroscience↗

bioRxiv2023

The A53T mutation in α-synuclein enhances pro-inflammatory activation in human microglia

Parkinsons disease (PD) is characterized by the aggregation of -synuclein into Lewy bodies and Lewy neurites in the brain. Microglia-driven neuroinflammation may contribute to neuronal death in PD, however the exact role of microglia remains unclear and has been understudied. The A53T mutation in the gene coding for -synuclein has been linked to early-onset PD, and exposure to A53T-mutant human -synuclein increases the potential for inflammation of murine microglia. To date, its effect has not been studied in human microglia. Here, we used 2-dimensional cultures of human iPSC-derived microglia and transplantation of these cells into the mouse brain to assess the effects of the A53T mutation on human microglia. We found that A53T-mutant human microglia had an intrinsically increased propensity towards pro-inflammatory activation upon inflammatory stimulus. Additionally, A53T mutant microglia showed a strong decrease in catalase expression in non-inflammatory conditions, and increased oxidative stress. Our results indicate that A53T mutant human microglia display cell-autonomous phenotypes that may worsen neuronal damage in early-onset PD.

neuroscience↗

bioRxiv2022

Cell projection plots: a novel visualization of bone marrow aspirate cytology

Deep models for cell detection have demonstrated utility in bone marrow cytology, showing impressive results in terms of accuracy and computational efficiency. However, these models have yet to be implemented in the clinical diagnostic workflow. Additionally, the metrics used to evaluate cell detection models are not necessarily aligned with clinical goals and targets. In order to address these issues, we introduce cell projection plots (CPPs), which are novel, automatically generated visual summaries of bone marrow aspirate specimens. CPPs provide a compact summary of bone marrow aspirate cytology, and encompass relevant biological patterns such as neutrophil maturation. To gauge clinical relevance, CPPs were shown to three hematopathologists, who decided whether shown diagnostic synopses matched with generated CPPs. Pathologists were able to match CPPs to the correct synopsis with 85% accuracy. Our finding suggests CPPs can compactly represent clinically relevant information from bone marrow aspirate specimens, and may be used to efficiently summarize bone marrow cytology to pathologists. CPP could be a step toward human-centered implementation of artificial intelligence (AI) in hematopathology, and a basis for a diagnostic support tool for digital pathology workflows.

pathology↗