Biology subjects

Zhang, L.

Publications and source records attributed to Zhang, L..

At least 559 records · Page 31Linked to original sources

bioRxiv2024

Relation of CMV and brain atrophy to trajectories of immunosenescence in diverse populations

Immunosenescence (ISC), the aging of the immune system, has largely been studied in populations of European descent. Here, circulating immune cell cytometric data from African-American, Hispanic, and non-Hispanic White participants were generated. Known and novel age effects were identified using either a meta-analysis approach or a parallel genetic approach. Most results are consistent across the three populations, but some cell populations display evidence of heterogeneity, such as a PD-L1+CD56+ NK cell subset. The study estimated "Immunological Age" (IA) during physiologic aging. While we found no relation of IA to Multiple Sclerosis, IA is associated with entorhinal cortex atrophy, a presymptomatic feature of Alzheimers disease, linking neurodegeneration and peripheral immunity. ISC trajectories were also inferred, highlighting age, CMV status, and genetic ancestry as key influences. Our assessment offers reference ISC trajectories for personalization of assessments of immune function over the life course in diverse populations.

immunology↗

bioRxiv2024

Role of R5 Pyocin in the Predominance of High-Risk Pseudomonas aeruginosa Isolates

Infections with antimicrobial resistant pathogens, such as Pseudomonas aeruginosa, are a frequent occurrence in healthcare settings. Human P. aeruginosa infections are predominantly caused by a small number of sequence types (ST), such as ST235, ST111, and ST175. Although ST111 is recognized as one of the most prevalent high-risk P. aeruginosa clones worldwide and frequently exhibits multidrug-resistant or extensively drug-resistant phenotypes, the basis for this dominance remains unclear. In this study, we used a genome-wide transposon insertion library screen to discover that the competitive advantage of ST111 strains over certain non-ST111 strains is through production of R pyocins. We confirmed this finding by showing that competitive dominance was lost by ST111 mutants with R pyocin gene deletions. Further investigation showed that sensitivity to ST111 R pyocin (specifically R5 pyocin) is caused by deficiency in the O-antigen ligase waaL, which leaves lipopolysaccharide (LPS) bereft of O antigen, enabling pyocins to bind the LPS core. In contrast, sensitivity of waaL mutants to R1 or R2 pyocins depended on additional genomic changes. In addition, we found the PA14 mutants in lipopolysaccharide biosynthesis (waaL, wbpL, wbpM) that cause high susceptibility to R pyocins also exhibit poor swimming motility. Analysis of 5,135 typed P. aeruginosa strains revealed that several international, high-risk sequence types (including ST235, ST111, and ST175) are enriched for R5 pyocin production, indicating a correlation between these phenotypes and suggesting a novel approach for evaluating risk from emerging prevalent P. aeruginosa strains. Overall, our study sheds light on the mechanisms underlying the dominance of ST111 strains and highlighting the role of waaL in extending spectrum of R pyocin susceptibility.

microbiology↗

bioRxiv2024

ProteinAligner: A Multi-modal Pretraining Framework for Protein Foundation Models

Protein foundation models, particularly protein language models, have demonstrated strong success in learning meaningful representations of proteins using transformer architectures pretrained on large-scale protein datasets with self-supervised learning. These representations have been highly effective for downstream tasks such as predicting protein functions and properties. However, most current protein foundation models focus on pretraining with amino acid sequences, often neglecting additional modalities like protein structures and related literature, both of which provide valuable insights. To address this gap, we propose a multi-modal pretraining approach that integrates three key modalities - protein sequences, structures, and literature text. In our framework, the protein sequence modality serves as the anchor, with the other two modalities aligned to it, enhancing the models capacity to capture more comprehensive protein information. ProteinAligner out-performed state-of-the-art protein foundation models in predicting protein functions and properties across diverse down-stream tasks.

bioinformatics↗

bioRxiv2024

Mitigation of multi-scale biases in cell-type deconvolution for spatially resolved transcriptomics using HarmoDecon

The advent of spatially resolved transcriptomics (SRT) has revolutionized our understanding of tissue molecular microenvironments by enabling the study of gene expression in its spatial context. However, many SRT platforms lack single-cell resolution, necessitating cell-type deconvolution methods to estimate cell-type proportions in SRT spots. Despite advancements in existing tools, these methods have not addressed biases occurring at three scales: individual spots, entire tissue samples, and discrepancies between SRT and reference scRNA-seq datasets. These biases result in overbalanced cell-type proportions for each spot, mismatched cell-type fractions at the sample level, and data distribution shifts across platforms. To mitigate these biases, we introduce HarmoDecon, a novel semi-supervised deep learning model for spatial cell-type deconvolution. HarmoDecon leverages pseudo-spots derived from scRNA-seq data and employs Gaussian Mixture Graph Convolutional Networks to address the aforementioned issues. Through extensive simulations on multi-cell spots from STARmap and osmFISH, HarmoDecon outperformed 11 state-of-the-art methods. Additionally, when applied to legacy SRT platforms and 10x Visium datasets, HarmoDecon achieved the highest accuracy in spatial domain clustering and maintained strong correlations between cancer marker genes and cancer cells in human breast cancer samples. These results highlight the utility of HarmoDecon in advancing spatial transcriptomics analysis.

bioinformatics↗

bioRxiv2024

Estimating descending activation patterns from EMG in fast and slow movements using a model of the stretch reflex

Due to spinal reflex loops, descending activation from the brain is not the only source of muscle activation that ultimately generates movement. This study directly estimates descending activation patterns from measured patterns of muscle activation (EMG) during human arm movements. A simple model of the spinal stretch reflex is calibrated in a postural unloading task and then used to estimate descending activation patterns from muscle EMG patterns and kinematics during voluntary arm motion performed at different speeds. We observed three key features of the estimated descending activation patterns: (1) Within about the first 15% of movement duration, descending and muscle activations are temporally aligned. Thereafter, they diverge and develop qualitatively different temporal profiles. (2) The time course of descending activation is monotonic for slow movements, non-monotonic for fast movements. (3) Varying model parameters like the spinal reflex gain or the level of co-contraction does not qualitatively change the temporal pattern of estimated descending activation. Our findings highlight the substantial contribution of spinal reflex loops to movement generation, while at the same time providing evidence that the brain must generate qualitatively different descending activation patterns for movements that vary in their mechanical dynamics. New & NoteworthyWe propose a new method that directly estimates descending activation from measured EMG signals and arm kinematics by inverting a model of the spinal stretch reflex, without the need for muscle models or for an arm dynamics model. This approach identifies key features of the time structure of descending activation as movement speed is varied, while also revealing the significant contribution of the spinal stretch reflex to movement generation.

neuroscience↗

bioRxiv2024

Multi-Modal Large Language Model Enables All-Purpose Prediction of Drug Mechanisms and Properties

Accurately predicting the mechanisms and properties of potential drug molecules is essential for advancing drug discovery. However, traditional methods often require the development of specialized models for each specific prediction task, resulting in inefficiencies in both model training and integration into work-flows. Moreover, these approaches are typically limited to predicting pharmaceutical attributes represented as discrete categories, and struggle with predicting complex attributes that are best described in free-form texts. To address these challenges, we introduce DrugChat, a multi-modal large language model (LLM) designed to provide comprehensive predictions of molecule mechanisms and properties within a unified framework. DrugChat analyzes the structure of an input molecule along with users queries to generate comprehensive, free-form predictions on drug indications, pharmacodynamics, and mechanisms of action. Moreover, DrugChat supports multi-turn dialogues with users, facilitating interactive and in-depth exploration of the same molecule. Our extensive evaluation, including assessments by human experts, demonstrates that DrugChat significantly outperforms GPT-4 and other leading LLMs in generating accurate free-form predictions, and exceeds state-of-the-art specialized prediction models.

pharmacology and toxicology↗

bioRxiv2024

Benchmarking copy number aberrations inference tools using single-cell multi-omics datasets

Copy number aberrations (CNAs) are an important type of genomic variation which play a crucial role in the initiation and progression of cancer. With the explosion of single-cell RNA sequencing (scRNA-seq), several computational methods have been developed to infer CNAs from scRNA-seq studies. However, to date, no independent studies have comprehensively benchmarked their performance. Herein, we evaluated five state-of-the-art methods based on their performance in tumor vs normal cell classification, CNAs profile accuracy, tumor subclone inference and aneuploidy identification in non-malignant cells. Our results showed that Numbat outperformed others across most evaluation criteria, while CopyKAT excelled in scenarios when expression matrix alone was used as input. In specific tasks, SCEVAN showed the best performance in clonal breakpoint detection and Numbat showed high sensitivity in copy number neutral LOH (cnLOH) detection. Additionally, we investigated how referencing settings, inclusion of tumor microenvironment cells, tumor type, and tumor purity impact the performance of these tools. This study provides a valuable guideline for researchers in selecting the appropriate methods for their datasets.

bioinformatics↗

bioRxiv2024

Combination of Imipridone ONC201 or ONC206 with Temozolomide and Radiotherapy in triple ITR therapy reduces intracranial tumor burden and prolongs survival in an orthotopic wild-type IDH GBM mouse model

Glioblastoma remains the most lethal common primary brain tumor in adults with limited therapeutic options. TIC10/ONC201, a first-in-class imipridone we discovered, achieved meaningful therapeutic effects in phase I/II trials in patients with diffuse gliomas harboring H3K27M mutations, and currently the drug is in randomized phase III testing (ACTION trial; NCT05580562). ONC201 targets mitochondrial protease ClpP to disrupt oxidative phosphorylation and trigger the integrated stress response (ISR), TRAIL/DR5, and tumor cell death. We hypothesized that ONC201 and its analogue ONC206 synergize with temozolomide (TMZ) and ionizing radiation (IR), standard-of-care glioblastoma therapies. ONC201 enhances TMZ or IR-induced apoptosis, and cytotoxicity. ClpP-silencing suppresses ONC201-induced cytotoxicity but not TMZ or RT. Both ONC201 and ONC206 reduce expression of TMZ-resistance mediator MGMT. Suppression of MGMT protein was observed in H3K27M-mutated DIPG cell lines following treatment with ONC201 or ONC206 with or without TMZ. Cytokine profiling indicates distinct ONC201 alterations relative to TMZ suggesting distinct anti-tumor immune mechanisms. Triple IR+TMZ+ONC201 (ITR) therapy prolongs median survival to 123 days with a tail on survival curve (3-of-7 mice alive beyond 200-days) in an orthotopic U251 GBM model versus ONC201 (44-days; p=0.000197), IR (63-days; p=0.0012), TMZ (78-days; p=0.0354), ONC201+IR (55-days; p=0.0004), ONC201+TMZ (80-days; p=0.0041) and IR+TMZ (103-days; p>0.05). By 231-days, the only surviving mice were in IRT group. Our results support investigation of ONC201/ONC206 in combination with TMZ and IR (ITR) in GBM or H3K27M mutated diffuse glioma therapy.

cancer biology↗

bioRxiv2024

FishCODE: a web-based information platform for comprehensive omics data exploration in fish research

In terms of the utilization of omics data, the current fish database analysis functions are primarily relatively simple tools at the transcriptional level, aimed at obtaining the co-expression levels of specified genes or the data visualization of multiple genes, and do not enable users to perform comprehensive omics data analysis. Furthermore, the gene-level information currently provided by these multispecies fish genomics databases is incomplete, and there is a lack of a comprehensive portal that can offer multidimensional genetic information. To address these challenges, we collected extensive multi-omics information on 35 fishes and established the primary comprehensive multi-omics data information platform for fish, FishCODE (http://bioinfo.ihb.ac.cn/fishcode). We have collected experimental background of dataset which pertaining to the target fishes, selected a range of datasets that encompass a broad spectrum of research areas, and downloaded the corresponding raw omics data from public repositories such as the Sequence Read Archive (SRA). Through a unified pipeline analysis, FishCODE contains 11,216 samples from 540 sets of genomic, transcriptomic, and methylomic datasets. These data encompass transcript structure and expression, gene methylation levels, protein domains, protein subcellular localization, protein interactions, best matched protein (Swiss-Prot), associated SNP site information (47,111,018), orthologous genes, phylogenetic tree and GO/KEGG annotations. To facilitate comparison, we annotated the experimental background data sets of the FishCODE, FishGET, PhyloFish, FishSED and FishSCT databases using the Fish Experimental Condition Ontology. Currently, the FishCODE database omics dataset includes 146 unique experimental condition words, 654 cumulative experimental condition words, and 13 species with rich experimental background (more than 20 unique FECO words). These data are 3.5 times (42), 8.3 times (74), and 6.5 times (2) those of the second-ranked databases respectively. We generated word cloud maps for the experimental condition vocabularies of FishCODE and FishGET, illustrating the superior richness of FishCODEs experimental background.

bioinformatics↗

bioRxiv2024

Balancing Growth: GROOT Genes Link Plant Biomass and Temperature Adaptation in Arabidopsis

Root systems take up water and nutrients from the soil and thereby underpin all essential plant functions. The size of the root system affects the ability of roots to capture these resources and determine the amount of carbon that roots transfer into the soil. Understanding the genetic basis of root biomass is therefore very important for enhancing crop resilience and productivity, especially in the face of climate change, as for soil carbon sequestration efforts. In our study, we catalogued root biomass of a diverse set of Arabidopsis accessions mainly derived from Sweden and Spain and utilized GWAS to identify loci associated with root biomass in Arabidopsis thaliana. We found that genetic variants associated with high biomass are enriched in accessions originating from distinct biogeographic regions in Spain. Among the most significant SNPs, one SNP on chromosome 3, was closely linked to three genes for which loss of function mutations caused significant increases in root, shoot and seed biomass and that we named GROOT genes. These genes act as general growth limiters. Additionally, our analysis revealed that the GROOT genes are in strong linkage disequilibrium, implying a potential coordinated function in regulating growth. Furthermore, accessions carrying the non-reference allele at this SNP showed markedly higher biomass under elevated temperatures, suggesting that these genes may also play a role in temperature adaptation. The discovery that the GROOT genes can enhance biomass without negative trade-offs across multiple traits opens new avenues for further research aimed at understanding genetically determined growth limitations, and for improving crop resilience and adaptability, as well as carbon sequestration.

plant biology↗

bioRxiv2024

Identification of Intrinsic Features for Cortical Separability of Human and Mouse Neurons

This work introduces a novel framework for holistic comparative analysis of cortical regions in mouse and human brains at single-neuron resolution, with a primary focus on the morphological and molecular characteristics of neurons. To do so, we generated one of the largest dendritic reconstruction datasets of cortical neurons to date, comprising 2,363 human neurons and 16,011 mouse neurons from the frontal, parietal, and temporal lobes, followed by establishing a rigorous procedure to identify anatomically and functionally corresponding brain regions with minimal variability in brain mapping. Additionally, we leveraged single nucleus/cell transcriptomic data from independent groups to validate the molecular correspondence of the brain regions identified in this study. The significance of these anatomically, functionally, and molecularly corresponding mouse-human region pairs is highlighted by examining the intrinsic features of their respective cortical regions. Our findings reveal that human neuron branching patterns differ dramatically from those in mouse brains, particularly in terms of dendritic branching frequency and normalized dendritic branching intervals. This difference is pronounced in the frontal and temporal lobes, underscoring the distinct neuronal architectures between the two species. At the single-neuron level, we found that neurons from the human frontal and parietal lobes are six times more separable than those from the same regions in mouse brains. This heightened separability is also observed between the frontal and temporal lobes, as well as between the parietal and temporal lobes in humans. We thoroughly explored the entire morphological feature space, along with its characteristic subspaces, and consistently found this distinct separability. Remarkably, this neuronal separability can be partially recapitulated when examining the global functional states of these brain lobes--using newly acquired Electroencephalography (EEG) and Magnetoencephalography (MEG) signals as physiological measures--as well as their global metabolic states, molecular profiles, and cortical geometry. These findings suggest that our comparative analysis of single-neuron intrinsic features could serve as a valuable foundation for future comprehensive studies of cross-species brain structures and functions.

neuroscience↗

bioRxiv2024

Microbial strategies of environmental adaptation revealed by trait-environmental relationships

Microbial trait variation along environment gradients is crucial to understanding their ecological adaptation mechanisms. With the increasing availability of microbial genomes, making full use of the genome-based traits to decipher their adaptation strategies becomes promising and urgent. Here, we examined microbial communities in water and sediments of 20 East African lakes with pH values ranging from 7.2 to 10.1 through taxonomic-profiling and genome-centric metagenomics. We identified functional traits important for microbial adaptation to the stresses of alkalinity and salinity based on the significant trait-environment relationships (TERs), including those involved in cytoplasmic pH homeostasis, compatible solute accumulation, cell envelope modification and energy requisition. By integrating these significant-TER traits, we further developed an environmental adaptation index to quantify the species-level adaptive capacity for environmental stresses, such as high pH environments. The adaptation index of pH showed consistently significant positive relationships with species pH optima across regional and global genomic datasets from freshwater, marine and soda lake ecosystems. The generality of the index for quantifying environmental adaptation was demonstrated by showing significant relationships with the species niche optima for the gradients of soil temperature and seawater salinity. These results highlight the importance of trait-environment relationships in facilitating the inference of microbial genomic-based adaptation mechanisms, and expand our understanding of ecological adaptative strategies along environmental gradients.

ecology↗

bioRxiv2024

Diphthamide formation in Arabidopsis requires DPH1-interacting DPH2 for light and oxidative stress resistance

Diphthamide is a post-translationally modified histidine residue of eukaryotic TRANSLATION ELONGATION FACTOR 2 (eEF2) and the target of diphtheria toxin in human cells. In yeast and mammals, the 4Fe-4S cluster-containing proteins Dph1 and Dph2 catalyze the first biosynthetic step of diphthamide formation. Here we identify Arabidopsis thaliana DPH2 and show that it is required for diphthamide biosynthesis, localizes to the cytosol and interacts physically with AtDPH1. Arabidopsis dph2 mutants form shorter primary roots and smaller rosettes than the wild type, similar to dph1 mutants which we characterized previously. Additionally, increased ribosomal -1 frameshifting error rates and attenuated TARGET OF RAPAMYCIN (TOR) kinase activity in dph2 mutants also phenocopy the dph1 mutant. Beyond the known heavy-metal hypersensitivity and heat shock tolerance of dph1, we newly show here that both dph1 and dph2 mutants are hypersensitive to elevated light intensities and oxidative stress, and that wild-type Arabidopsis seedlings accumulate diphthamide-unmodified eEF2 under oxidative stress. Both mutants share the deregulation of 1,186 transcripts in numerous environmental and hormone responses. AtDPH1 and AtDPH2 do not complement the corresponding mutants of Saccharomyces cerevisiae. In summary, DPH2 and DPH1 interact to function inter-dependently in diphthamide formation, the maintenance of translational fidelity, wild-type growth rates and TOR kinase activation, and they contribute to mitigating damage from elevated light intensities and oxidative stress. Under oxidative stress, a dose-dependent loss of diphthamide could potentiate downstream effects in a feed-forward loop. This work advances our understanding of translation and its interactions with growth regulation and stress responses in plants.

plant biology↗

bioRxiv2024

An integrated view of the structure and function of the human 4D nucleome

The dynamic three-dimensional (3D) organization of the human genome (the "4D Nucleome") is closely linked to genome function. Here, we integrate a wide variety of genomic data generated by the 4D Nucleome Project to provide a detailed view of human 3D genome organization in widely used embryonic stem cells (H1-hESCs) and immortalized fibroblasts (HFFc6). We provide extensive benchmarking of 3D genome mapping assays and integrate these diverse datasets to annotate spatial genomic features across scales. The data reveal a rich complexity of chromatin domains and their sub-nuclear positions, and over one hundred thousand structural loops and promoter-enhancer interactions. We developed 3D models of population-based and individual cell-to-cell variation in genome structure, establishing connections between chromosome folding, nuclear organization, chromatin looping, gene transcription, and DNA replication. We demonstrate the use of computational methods to predict genome folding from DNA sequence, uncovering potential effects of genetic variants on genome structure and function. Together, this comprehensive analysis contributes insights into human genome organization and enhances our understanding of connections between the regulation of genome function and 3D genome organization in general.

genomics↗

bioRxiv2024

Rapid production of modified cattle and sheep from spermatid-like stem cells

Genome-modified ruminants, such as cattle and sheep, are important germplasm for agricultural applications. Haploid androgenetic stem cells (haSCs) can support the full-term development of embryos upon intracytoplasmic haSCs injection (iCHI) into intact oocytes. Thus, haSCs are invaluable resources for studying economic traits and significantly impact livestock breeding; however, ruminant haSCs have yet to be obtained. Here, we report the derivation of cattle and sheep haSCs using a recombinant FACE medium. These cells display characteristics of formative-state pluripotency and can differentiate into three germ layers, both in vitro and in vivo. Interestingly, the ectopic expression of protamine in haSCs converts their nuclei into a spermatid-like structure and synergistic enhancement of full-term development of iCHI embryos, named as Pro-iCHI approach. Furthermore, gene-modified cattle and sheep can be efficiently produced through the combinatorial use of the CRISPR and Pro-iCHI approach. Overall, Pro-iCHI provides an easy-to-manipulate agrobiotechnology tool for rapidly producing genetically modified livestock.

bioengineering↗

bioRxiv2024

PSMA-specific CAR-engineered macrophages for therapy of prostate cancer

Chimeric antigen receptor (CAR)-modified macrophages (CAR-Ms) are a promising approach for the treatment of solid tumors due to its high infiltration and immune-regulation activity. Prostate cancer is a typical solid tumor associated with highly immunosuppressive microenvironment. To date, the potential application of CAR-M cell therapy in prostate cancer has been infrequently explored. The prostate-specific membrane antigen (PSMA) functions as a specific biomarker for prostate cancer. In this study, we assessed the antitumor efficacy of PSMA-targeted CAR-Ms in preclinical models. CAR-Ms were engineered to express a PSMA-specific single-chain variable fragment (scFv) and co-stimulatory domains. In vitro data demonstrated specific cytotoxicity of CAR-Ms against PSMA-expressing prostate cancer cells, which was further supported by transcriptome analysis demonstrating the pro-inflammatory phenotypes of CAR-Ms. In vivo studies using xenograft mouse models confirmed significant tumor regression after administration of PSMA-targeted CAR-Ms compared to controls. Histopathological analysis showed infiltration of CAR-Ms into tumor tissues without off-target toxicity. These results highlight the strong antitumor activity and safety of PSMA-targeted CAR-Ms, supporting their potential as a new immunotherapy for prostate cancer.

bioengineering↗

bioRxiv2024

Multi-ancestry GWAS reveals loci linked to human variation in LINE-1- and Alu-copy numbers

LINE-1 (L1) and Alu are two families of transposable elements (TEs) occupying [~]17% and [~]11% of the human genome, respectively. Though only a small fraction of L1 copies is able to produce the machinery to mobilize autonomously, Alu and degenerate L1s can hijack their functional machinery and mobilize in trans. The expression and subsequent mobilization of L1 and Alu can exert pathological effects on their hosts. These features have made them promising focus subjects in studies of aging where they can become active. However, mechanisms regulating TE activity are incompletely characterized, especially in diverse human populations. To address these gaps, we leveraged genomic data from the 1000 Genomes Project to carry out a trans-ethnic GWAS of L1/Alu insertion singletons. These are rare, recently acquired insertions observed in only one person and which we used as proxies for variation in L1/Alu insertion numbers. Our approach identified SNVs in genomic regions containing genes with potential and known TE regulatory properties, and it enriched for SNVs in regions containing known regulators of L1 expression. Moreover, we identified reference TE copies and structural variants that associated with L1/Alu singletons, suggesting their potential contribution to TE insertion number variation. Finally, a transcriptional analysis of lymphoblastoid cells highlighted potential cell cycle alterations in a subset of samples harboring L1/Alu singletons. Collectively, our results suggest that known TE regulatory mechanisms may be active in diverse human populations, expand the list of loci implicated in TE insertion number variability, and reinforce links between TEs and disease.

genomics↗

bioRxiv2024

Numerosity adaptation suppresses early visual responses

Humans and many animals rapidly and accurately perceive numerosity, the number of objects, in a visual image. The numerosity of recently viewed images influences our perception of the current images numerosity: numerosity adaptation. How does numerosity adaptation affect responses to numerosity in the brain? Recent studies show both early visual responses that monotonically increase with numerosity, and later numerosity-tuned responses that peak at different (preferred) numerosities in different neural populations. We have recently shown that numerosity adaptation affects the preferred numerosity of numerosity-tuned neural populations. We have also shown that early visual monotonic responses reflect image contrast, which follows numerosity closely. Here we ask how monotonic responses in the early visual cortex are affected by adaptation to different numerosities, using ultra-high field (7T) fMRI and neural model-based analyses. FMRI response amplitudes increased monotonically with numerosity throughout the early visual field maps (V1-V3, hV4, LO1-LO2 & V3A/B). This increase in response amplitudes becomes less steep after adaptation to higher numerosities, with this effect becoming stronger through the early visual hierarchy. This suppression of responses to numerosity is consistent with perceptual effects where adaptation to high numerosities reduces the perceived numerosity. These results imply that numerosity adaptation effects in later numerosity-tuned neural populations may originate in early visual areas that respond to image contrast in the adapting image.

neuroscience↗