The Lieber Institute's comprehensive integration of single cell and spatial transcriptomics redefines our understanding of NAc architecture, and opens new avenues for translational neuroscience
The nucleus accumbens (NAc) has occupied a central position in reward neuroscience for decades. Its GABAergic medium spiny neurons (MSNs), dopaminergic inputs from the ventral tegmental area, and glutamatergic afferents from cortical and limbic structures have been extensively characterized in rodent models. But the degree to which that rich body of rodent data translates to human NAc biology has remained an open and consequential question. Species differences in anatomy, cell type composition, and transcriptional organization are not trivial, and for a brain region as clinically relevant as the NAc, the gap between rodent and human data carries real implications for understanding psychiatric disease and developing treatments.
A new preprint from Ravichandran, Bach, Phillips, and colleagues at the Lieber Institute for Brain Development addresses this gap directly, delivering the first integrated spatiomolecular atlas of the human NAc. By pairing spatial transcriptomics with single-nucleus RNA sequencing (snRNA-seq) across postmortem tissue from ten neurotypical adult donors, the team has produced a resource that simultaneously characterizes cellular diversity, spatial organization, cross-species conservation, disease-relevant signaling, and drug-responsive transcriptional programs, all within a unified analytical framework.
Building the Atlas
Using 10x Genomics Visium and Chromium platforms on adjacent tissue sections, the authors generated paired spatially-resolved transcriptomics (SRT) and snRNA-seq data from NAc tissue dissected to capture the complete mediolateral and dorsoventral extent of the structure.
After rigorous quality control, the snRNA-seq dataset retained 103,785 high-quality nuclei, yielding 20 transcriptionally distinct cell populations: 13 neuronal and 7 non-neuronal. As expected, the majority of neurons represented GABAergic MSNs expressing dopamine receptor 1 (DRD1) or 2 (DRD2) . The authors identified four DRD1 MSN subtypes and two DRD2 MSN subtypes, each with distinct transcriptional profiles. Of particular interest, two DRD1 MSN subtypes showed strong transcriptional similarity to previously described DRD1 MSN populations in rodents and non-human primates that form D1 islands along the ventromedial NAc border . The remaining neuronal diversity included cholinergic neurons, parvalbumin neurons, somatostatin/cortistatin neurons, VIP neurons, GLP1R+ GABAergic neurons, and a previously undescribed inhibitory population marked by KCNC2 and ANK1.
On the spatial side, 176,013 high-quality Visium spots were clustered, yielding eight spatial domains: MSN_1, MSN_2, MSN_3, D1 islands, Inhibitory, Excitatory, Endothelial/Ependymal, and White Matter. Further analyses identified that MSN domains are characterized by transcriptional gradients consisting of multiple MSN subtypes. Integration of genome-wide association (GWAS) summary statistics with cell-type and domain marker genes linked spatial domains and transcriptional gradients to genetic risk for psychiatric disorders and addictions-related trains.
The D1 Islands: A Conserved and Functionally Heterogeneous Compartment
One of the study’s most compelling contributions is its characterization of D1 interface islands, specialized MSN clusters along the NAc border that have been described in rodents and NHPs but never directly characterized in the human brain using molecular tools.
The authors identify enrichment of OPRM1 encoding the mu opioid receptor in human D1 islands, consistent with prior cross-species observations. Critically, they identify three transcriptionally distinct gene expression patterns within D1 islands (nmf34, nmf35, nmf44), each with unique spatial distributions and molecular signatures. nmf34, restricted to ventromedial D1 islands, correlates strongly with CPNE4 and TAC1 — markers of DRD1_MSN_B. nmf44 shows the strongest enrichment for depression risk, while all three factors are enriched for schizophrenia and bipolar disorder heritability. The identification of intra-island heterogeneity, and its differential association with psychiatric genetic risk, is a finding with significant implications for understanding how opioid and neuromodulatory signaling in the NAc contributes to affective disorders and addiction.
Ligand-receptor analysis further reinforces the functional significance of D1 islands, showing convergent opioid signaling through both PENK-OPRM1 and PDYN-OPRM1 pathways onto DRD1_MSN_B and DRD1_MSN_D populations localized to D1 islands. CRH-CRHR1 signaling through specific inhibitoryneurons onto DRD1_MSN_D adds a stress-axis dimension consistent with the proposed role of these islands in integrating hedonic and aversive signals.
Why This Resource Matters for the Field
Several aspects of this dataset make it particularly valuable for the broader neuroscience community. First, the study provides a roadmap for integrating single cell and spatial transcriptomic data using factorization approaches for resolving continuous transcriptional gradients in brain structures where discrete clustering approaches miss biologically meaningful variation. The identification of medial-lateral transcriptional gradients in the human NAc — rather than a sharp core/shell dichotomy — has direct implications for understanding NAc function in the human brain.
Second, the cross-species registration framework, correlating human spatial domain and cell type signatures against both NHP and rodent references, provides a comprehensive resource for evaluating evolutionary conservation and translational validity of animal models. Third, the drug-response mapping approach, projecting rodent-derived transcriptional signatures associated with morphine and cocaine exposure onto human SRT data, establishes a computational framework that can be applied broadly across other snRNA-seq and SRT datasets to spatially localize transcriptional effects of pharmacological and environmental perturbations.
Finally, the study associates specific NAc cell types and spatial domains with genetic risk for neuropsychiatric disorders and addiction-related traits.
Future Directions
Several high-priority extensions follow naturally from this work. Systematic sampling along the full anterior-posterior axis within individual donors is needed to fully characterize NAc transcriptional gradients and spatial features, particularly given the authors’ observation that D1 islands are prominent in anterior and intermediate sections but largely absent posteriorly. Higher-resolution imaging-based SRT platforms, such as VisiumHD or Xenium, would enable precise co-localization of cell types within the spatial gradients identified here, resolving the intermixing of DRD1 and DRD2 MSNs that current Visium resolution cannot fully separate.
Case-control datasets from individuals with substance use disorders, depression, or schizophrenia, profiled with paired snRNA-seq and SRT, represent the most direct next step for translating this atlas into mechanistic and therapeutic insight.
Conclusion: Embracing Regulatory Complexity
This spatiomolecular atlas of the human NAc is a foundational resource for the fields of neuroscience, psychiatric genetics, and addiction . It establishes that the human NAc contains substantially more cellular and spatial complexity than prior models suggested, demonstrates cross-species conservation of key organizational features while clarifying important divergences, and provides a spatially resolved framework for interpreting genetic risk for neuropsychiatric traits. The dataset, tools, and analytical approaches introduced here are openly available and positioned to accelerate research across multiple fields.
We’ve created an accompanying article that focuses on the big picture and real-world impact of this research, without the technical details.
Dr. Kristen Maynard leads the Molecular Neuroanatomy Team at the Lieber Institute for Brain Development. Her work focuses on exploring gene expression, brain structure, and brain function in relation to psychiatric diseases.