Evan Collins
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Evan Collins has a total of 66 co-authors. Here are the top co-authors by number of shared publications:
- Hitten P. Zaveri
Departments of Neurology, Yale University
4 shared publications - Adithya Sivaraju
Yale New Haven Hospital Yale University New Haven Connecticut USA
3 shared publications - Daniel G. Anderson
David H. Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology, Cambridge, USA
3 shared publications - Robert Langer
David H. Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology, Cambridge, USA
3 shared publications Alex King
Department of Neurosurgery, Yale School of Medicine, New Haven, CT, United States
2 shared publicationsDennis D. Spencer
Department of Neurosurgery, Yale School of Medicine, New Haven, CT, United States
2 shared publications
Latest Publications
Language mapped to a high‐resolution brain atlas for surgical evaluation of epilepsy patients
Objective: We created composite maps of language function from extraoperative stimulation literature and transformed them to the Yale Brain Atlas (YBA), which offers precise cortical localization with 690 one cm 2 parcels, based on the MNI152 template and anatomical landmarks. This allowed comparison to similarly transformed direct cortical stimulation (DCS) maps created from medically intractable epilepsy patients studied intracranially at Yale University and selected fMRI activation data. Our goal was to create anatomically precise boundaries of language function and support individualized planning for intracranial EEG (icEEG) studies and/or surgical resection. Methods: A systematic search of DCS studies of language‐related functions identified 12 stimulation studies (10 subdural, two stereo‐EEG) that provided sufficiently clear surface visualizations for parcel‐level mapping. Language‐positive sites from six commonly tested functions (visual naming, auditory naming, naming (auditory/visual), reading, repetition and speech arrest) were manually co‐registered to YBA parcels using conserved anatomical landmarks, facilitating topological comparison among similar language functions from the literature, Yale's electrical stimulation mapping cohort, and Neurosynth and Parcelsynth fMRI activation. Results: Across 1182 patients, 176 YBA parcels (approximately 25% of cortical parcels) were implicated in at least one language task. Composite stimulation maps demonstrated robust consistency with the Yale DCS cohort and substantial overlap with distributed fMRI language networks. The resulting interactive maps provide parcel‐level specificity while highlighting under‐sampled but functionally critical regions. Significance: The maps provide a review of language function determined through DCS and meta‐analytic repositories of fMRI activation data. By consolidating stimulation‐derived language data into a high‐resolution parcellated atlas, this work offers a spatially detailed surface‐based representation of language boundaries for surgical planning. The YBA facilitates standardized multimodal communication and allows patient‐specific stimulation results to be interpreted within population‐derived functional boundaries.
Language mapped to a high‐resolution brain atlas for surgical evaluation of epilepsy patients
Objective: We created composite maps of language function from extraoperative stimulation literature and transformed them to the Yale Brain Atlas (YBA), which offers precise cortical localization with 690 one cm 2 parcels, based on the MNI152 template and anatomical landmarks. This allowed comparison to similarly transformed direct cortical stimulation (DCS) maps created from medically intractable epilepsy patients studied intracranially at Yale University and selected fMRI activation data. Our goal was to create anatomically precise boundaries of language function and support individualized planning for intracranial EEG (icEEG) studies and/or surgical resection. Methods: A systematic search of DCS studies of language‐related functions identified 12 stimulation studies (10 subdural, two stereo‐EEG) that provided sufficiently clear surface visualizations for parcel‐level mapping. Language‐positive sites from six commonly tested functions (visual naming, auditory naming, naming (auditory/visual), reading, repetition and speech arrest) were manually co‐registered to YBA parcels using conserved anatomical landmarks, facilitating topological comparison among similar language functions from the literature, Yale's electrical stimulation mapping cohort, and Neurosynth and Parcelsynth fMRI activation. Results: Across 1182 patients, 176 YBA parcels (approximately 25% of cortical parcels) were implicated in at least one language task. Composite stimulation maps demonstrated robust consistency with the Yale DCS cohort and substantial overlap with distributed fMRI language networks. The resulting interactive maps provide parcel‐level specificity while highlighting under‐sampled but functionally critical regions. Significance: The maps provide a review of language function determined through DCS and meta‐analytic repositories of fMRI activation data. By consolidating stimulation‐derived language data into a high‐resolution parcellated atlas, this work offers a spatially detailed surface‐based representation of language boundaries for surgical planning. The YBA facilitates standardized multimodal communication and allows patient‐specific stimulation results to be interpreted within population‐derived functional boundaries.
- Top Cited PapersOpen Access
- 28 January 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Lipid Nanoparticle Database towards structure-function modeling and data-driven design for nucleic acid delivery
Lipid nanoparticles (LNPs) are the leading nonviral nucleic acid delivery technology, but LNP structure-function data remains fragmented and nonstandardized. Unlike protein engineering which is anchored by the centralized Protein Data Bank, the LNP field lacks a unified repository for systematic analysis. To address this, we develop Lipid Nanoparticle Database (LNPDB) ( https://lnpdb.molcube.com ), an integrated database and web tool that consolidates structural and functional data for 19,528 LNPs. LNPDB standardizes LNP featurization by encoding lipid composition, experimental methods, and functional results, and generates CHARMM force field files for constituent lipids to enable molecular dynamics simulations. LNPDB also supports future data contributions for continued growth. We examine the utility of LNPDB through two applications: advancing our deep learning model for predicting LNP delivery performance, and simulating bilayer dynamics to identify structural features – bilayer stability and critical packing parameter – that correlate with LNP delivery performance. Altogether, LNPDB provides the digital framework for LNP modeling and data-driven rational design.
- 28 January 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Top Cited PapersOpen AccessLipid Nanoparticle Database towards structure-function modeling and data-driven design for nucleic acid delivery
Lipid nanoparticles (LNPs) are the leading nonviral nucleic acid delivery technology, but LNP structure-function data remains fragmented and nonstandardized. Unlike protein engineering which is anchored by the centralized Protein Data Bank, the LNP field lacks a unified repository for systematic analysis. To address this, we develop Lipid Nanoparticle Database (LNPDB) ( https://lnpdb.molcube.com ), an integrated database and web tool that consolidates structural and functional data for 19,528 LNPs. LNPDB standardizes LNP featurization by encoding lipid composition, experimental methods, and functional results, and generates CHARMM force field files for constituent lipids to enable molecular dynamics simulations. LNPDB also supports future data contributions for continued growth. We examine the utility of LNPDB through two applications: advancing our deep learning model for predicting LNP delivery performance, and simulating bilayer dynamics to identify structural features – bilayer stability and critical packing parameter – that correlate with LNP delivery performance. Altogether, LNPDB provides the digital framework for LNP modeling and data-driven rational design.
- 11 November 2025
- Published by Wolters Kluwer Health in Neurology
- Vol. 105 (9) , e214256
Multitask Language Mapping to Visualize the Spatial Configuration of Polyfunctional Language Cortex
Direct electrical stimulation (DES) remains the gold standard for identifying eloquent language cortex. However, most protocols rely on single-task mapping, typically visual naming, potentially overlooking modality-specific or distributed language areas. The aim of this study was to define the spatial extent of polyfunctional language cortex using multitask DES integrated with the Yale Brain Atlas (YBA) and to supplement this with population-based structural connectivity data. We retrospectively analyzed patients with drug-resistant focal epilepsy who underwent intracranial EEG monitoring and extraoperative DES language mapping at Yale (2017-2022). Six tasks were administered at each bipolar contact pair: visual naming, auditory naming, reading, repetition, auditory comprehension, and written comprehension. Responses were mapped to the YBA (∼1-cm2 resolution), normalized for amperage and sampling frequency, and aggregated across patients. Intertask correlations were assessed using Kendall rank coefficients. Structural connectivity between language-involved parcels was evaluated using deterministic tractography from diffusion data of 1,065 healthy participants. Analyses were conducted using Python 3.9.13. Fifteen patients were included (mean age, 42.8 ± 14.1 years; 53% male). Multitask mapping identified a reproducible "language core" involving the superior temporal gyrus, pars triangularis, anterior supramarginal gyrus, and a basal temporal area located 4-6 cm posterior to the temporal tip. A tiered task strategy-visual naming, auditory naming, and reading-captured over 99% of all language-positive sites. At the individual level, up to 16% (8.2%-15.9%) of parcels cleared by visual naming alone exhibited disruption with other tasks, notably reading. Intertask correlation was highest between auditory naming and visual naming (p < 0.0001, τ = 0.70, 95% CI 0.61-0.78). Structural connectivity analysis showed involvement of the arcuate fasciculus (76.9%), parietal aslant tract (85.4%), and superior longitudinal fasciculus (73.9%). The supramarginal region had the densest language connectivity (69.2%), although connectivity strength did not correlate with DES-defined language involvement (τ = -0.14, p = 0.11). Multitask DES anchored to the YBA provides an anatomically precise, reproducible platform for language mapping. A focused task set enhances efficiency without sacrificing sensitivity. Integration with structural connectivity offers network-level insights. Limitations include modest sample size, variable electrode coverage with undersampling of the anterior language areas, and patient population with epilepsy, which may affect generalizability. This multimodal framework informs surgical planning and advances our understanding of human language representation.
- 11 November 2025
- Published by Wolters Kluwer Health in Neurology
- Vol. 105 (9) , e214256
Multitask Language Mapping to Visualize the Spatial Configuration of Polyfunctional Language Cortex
Direct electrical stimulation (DES) remains the gold standard for identifying eloquent language cortex. However, most protocols rely on single-task mapping, typically visual naming, potentially overlooking modality-specific or distributed language areas. The aim of this study was to define the spatial extent of polyfunctional language cortex using multitask DES integrated with the Yale Brain Atlas (YBA) and to supplement this with population-based structural connectivity data. We retrospectively analyzed patients with drug-resistant focal epilepsy who underwent intracranial EEG monitoring and extraoperative DES language mapping at Yale (2017-2022). Six tasks were administered at each bipolar contact pair: visual naming, auditory naming, reading, repetition, auditory comprehension, and written comprehension. Responses were mapped to the YBA (∼1-cm2 resolution), normalized for amperage and sampling frequency, and aggregated across patients. Intertask correlations were assessed using Kendall rank coefficients. Structural connectivity between language-involved parcels was evaluated using deterministic tractography from diffusion data of 1,065 healthy participants. Analyses were conducted using Python 3.9.13. Fifteen patients were included (mean age, 42.8 ± 14.1 years; 53% male). Multitask mapping identified a reproducible "language core" involving the superior temporal gyrus, pars triangularis, anterior supramarginal gyrus, and a basal temporal area located 4-6 cm posterior to the temporal tip. A tiered task strategy-visual naming, auditory naming, and reading-captured over 99% of all language-positive sites. At the individual level, up to 16% (8.2%-15.9%) of parcels cleared by visual naming alone exhibited disruption with other tasks, notably reading. Intertask correlation was highest between auditory naming and visual naming (p < 0.0001, τ = 0.70, 95% CI 0.61-0.78). Structural connectivity analysis showed involvement of the arcuate fasciculus (76.9%), parietal aslant tract (85.4%), and superior longitudinal fasciculus (73.9%). The supramarginal region had the densest language connectivity (69.2%), although connectivity strength did not correlate with DES-defined language involvement (τ = -0.14, p = 0.11). Multitask DES anchored to the YBA provides an anatomically precise, reproducible platform for language mapping. A focused task set enhances efficiency without sacrificing sensitivity. Integration with structural connectivity offers network-level insights. Limitations include modest sample size, variable electrode coverage with undersampling of the anterior language areas, and patient population with epilepsy, which may affect generalizability. This multimodal framework informs surgical planning and advances our understanding of human language representation.
- 1 October 2025
- Published by Springer Nature in Nature Computational Science
- Vol. 5 (11) , 976-979
Self-driving labs for biotechnology
- 1 October 2025
- Published by Springer Nature in Nature Computational Science
- Vol. 5 (11) , 976-979
Self-driving labs for biotechnology
- Open Access
- 18 June 2025
- Published by Frontiers Media SA in Frontiers in Network Physiology
- Vol. 5, 1585019
Yale Brain Atlas to interactively explore multimodal structural and functional neuroimaging data
Understanding the relationship between structure and function in the human brain is essential for revealing how brain organization influences cognition, perception, emotion, and behavior. To this end, we introduce an interactive web tool and underlying database for Yale Brain Atlas, a high-resolution anatomical parcellation designed to facilitate precise localization and generalizable analyses of multimodal neuroimaging data. The tool supports parcel-level exploration of structural and functional data through dedicated interactive pages for each modality. For structural data, it incorporates white matter connectomes of 1,065 subjects and cortical thickness profiles of 200 subjects both from the Human Connectome Project. For functional data, it includes resting-state fMRI connectivity matrices for 34 healthy subjects and task-specific fMRI activation data acquired from two meta-analytic resources–Neurosynth and NeuroQuery–which, once translated into Yale Brain Atlas space and modified to include 334 function-specific terms, form Parcelsynth and ParcelQuery, respectively. Altogether, to support investigation of brain structure-function relationships, this study presents a web tool and database for the Yale Brain Atlas that enable scalable, interactive exploration of multimodal neuroimaging data.
- 18 June 2025
- Published by Frontiers Media SA in Frontiers in Network Physiology
- Vol. 5, 1585019
Open AccessYale Brain Atlas to interactively explore multimodal structural and functional neuroimaging data
Understanding the relationship between structure and function in the human brain is essential for revealing how brain organization influences cognition, perception, emotion, and behavior. To this end, we introduce an interactive web tool and underlying database for Yale Brain Atlas, a high-resolution anatomical parcellation designed to facilitate precise localization and generalizable analyses of multimodal neuroimaging data. The tool supports parcel-level exploration of structural and functional data through dedicated interactive pages for each modality. For structural data, it incorporates white matter connectomes of 1,065 subjects and cortical thickness profiles of 200 subjects both from the Human Connectome Project. For functional data, it includes resting-state fMRI connectivity matrices for 34 healthy subjects and task-specific fMRI activation data acquired from two meta-analytic resources–Neurosynth and NeuroQuery–which, once translated into Yale Brain Atlas space and modified to include 334 function-specific terms, form Parcelsynth and ParcelQuery, respectively. Altogether, to support investigation of brain structure-function relationships, this study presents a web tool and database for the Yale Brain Atlas that enable scalable, interactive exploration of multimodal neuroimaging data.
- Top Cited Papers
- 10 December 2024
- Published by Springer Nature in Nature Biotechnology
- Vol. 43 (11) , 1790-1799
Artificial intelligence-guided design of lipid nanoparticles for pulmonary gene therapy
Ionizable lipids are a key component of lipid nanoparticles, the leading nonviral messenger RNA delivery technology. Here, to advance the identification of ionizable lipids beyond current methods, which rely on experimental screening and/or rational design, we introduce lipid optimization using neural networks, a deep-learning strategy for ionizable lipid design. We created a dataset of >9,000 lipid nanoparticle activity measurements and used it to train a directed message-passing neural network for prediction of nucleic acid delivery with diverse lipid structures. Lipid optimization using neural networks predicted RNA delivery in vitro and in vivo and extrapolated to structures divergent from the training set. We evaluated 1.6 million lipids in silico and identified two structures, FO-32 and FO-35, with local mRNA delivery to the mouse muscle and nasal mucosa. FO-32 matched the state of the art for nebulized mRNA delivery to the mouse lung, and both FO-32 and FO-35 efficiently delivered mRNA to ferret lungs. Overall, this work shows the utility of deep learning for improving nanoparticle delivery.
- 10 December 2024
- Published by Springer Nature in Nature Biotechnology
- Vol. 43 (11) , 1790-1799
Top Cited PapersArtificial intelligence-guided design of lipid nanoparticles for pulmonary gene therapy
Ionizable lipids are a key component of lipid nanoparticles, the leading nonviral messenger RNA delivery technology. Here, to advance the identification of ionizable lipids beyond current methods, which rely on experimental screening and/or rational design, we introduce lipid optimization using neural networks, a deep-learning strategy for ionizable lipid design. We created a dataset of >9,000 lipid nanoparticle activity measurements and used it to train a directed message-passing neural network for prediction of nucleic acid delivery with diverse lipid structures. Lipid optimization using neural networks predicted RNA delivery in vitro and in vivo and extrapolated to structures divergent from the training set. We evaluated 1.6 million lipids in silico and identified two structures, FO-32 and FO-35, with local mRNA delivery to the mouse muscle and nasal mucosa. FO-32 matched the state of the art for nebulized mRNA delivery to the mouse lung, and both FO-32 and FO-35 efficiently delivered mRNA to ferret lungs. Overall, this work shows the utility of deep learning for improving nanoparticle delivery.
- Open Access
- 16 August 2024
- Published by Springer Nature in Nature Communications
- Vol. 15 (1) , 1-15
Mapping the structure-function relationship along macroscale gradients in the human brain
Functional coactivation between human brain regions is partly explained by white matter connections; however, how the structure-function relationship varies by function remains unclear. Here, we reference large data repositories to compute maps of structure-function correspondence across hundreds of specific functions and brain regions. We use natural language processing to accurately predict structure-function correspondence for specific functions and to identify macroscale gradients across the brain that correlate with structure-function correspondence as well as cortical thickness. Our findings suggest structure-function correspondence unfolds along a sensory-fugal organizational axis, with higher correspondence in primary sensory and motor cortex for perceptual and motor functions, and lower correspondence in association cortex for cognitive functions. Our study bridges neuroscience and natural language to describe how structure-function coupling varies by region and function in the brain, offering insight into the diversity and evolution of neural network properties.
- 16 August 2024
- Published by Springer Nature in Nature Communications
- Vol. 15 (1) , 1-15
Open AccessMapping the structure-function relationship along macroscale gradients in the human brain
Functional coactivation between human brain regions is partly explained by white matter connections; however, how the structure-function relationship varies by function remains unclear. Here, we reference large data repositories to compute maps of structure-function correspondence across hundreds of specific functions and brain regions. We use natural language processing to accurately predict structure-function correspondence for specific functions and to identify macroscale gradients across the brain that correlate with structure-function correspondence as well as cortical thickness. Our findings suggest structure-function correspondence unfolds along a sensory-fugal organizational axis, with higher correspondence in primary sensory and motor cortex for perceptual and motor functions, and lower correspondence in association cortex for cognitive functions. Our study bridges neuroscience and natural language to describe how structure-function coupling varies by region and function in the brain, offering insight into the diversity and evolution of neural network properties.
- Open Access
- 13 March 2024
- Published by Elsevier in Brain Stimulation
- Vol. 17 (2) , 339-345
Systematic 1 Hz direct electrical stimulation for seizure induction: A reliable method for localizing seizure onset zone and predicting seizure freedom
- 13 March 2024
- Published by Elsevier in Brain Stimulation
- Vol. 17 (2) , 339-345
Open AccessSystematic 1 Hz direct electrical stimulation for seizure induction: A reliable method for localizing seizure onset zone and predicting seizure freedom