choucavalier
commented
4 years ago @jsheunis is it possible to remove this submission? I accidentally submitted it in its non-final form. thanks!
Open jsheunis opened 4 years ago
choucavalier
commented
4 years ago @jsheunis is it possible to remove this submission? I accidentally submitted it in its non-final form. thanks!
jsheunis
commented
4 years ago No worries. Maybe you can send me the updated version here in a comment, then I'll update the content above?
choucavalier
commented
4 years ago Works for me! Thanks š
choucavalier
commented
4 years ago @jsheunis thank you very much for accepting that I update the submission!
I would like to make 2 changes:
**Presentors**
- Gaston Zanitti (@gzanitti) & Valentin Iovene (@tgy)
**Collaborators**
- Gaston Zanitti (@gzanitti) & Valentin Iovene (@tgy)
- Antonia Machlouzarides-Shalit
- Nikita Zdanovitch (@Jbat1Jumper)
- Demian Wassermann (@demianw)
**Affiliations**: @Parietal-INRIA, @neurospin
**Abstract**
Coordinate-based meta-analysis (CBMA) databases like [Neurosynth](https://github.com/neurosynth) are successfully used within brain mapping studies to define regions of interest supported by past literature. To the best of our knowledge, no tool is currently available to the brain mapping community for conveniently incorporating a priori neuroanatomical and ontological knowledge into richer meta-analyses. Through a language-oriented programming methodological approach, standing on decades of computer science research on logic programming languages and their probabilistic extensions, we build NeuroLang: a language for expressing rich multi-modal brain mapping hypotheses combining CBMA databases with neuroanatomical atlases (e.g. Destrieux et al.) and ontological knowledge bases (e.g. NeuroFMA).
In NeuroLang, CBMA databases, neuroanatomical atlases and ontological knowledge are seamlessly combined through declarative logic rules, forming a program. These rules are extended with probabilistic semantics to model the uncertainty inherent to noisy neuroimaging data. In all, NeuroLang is a fully-fledged probabilistic logic programming language. For instance, querying a NeuroLang program can produce forward inference maps that account for all the knowledge necessary to answer the query.
This demo walks you through examples of rich forward inference brain maps obtained from NeuroLang programs. We produce a brain map of auditory cognitive processes from a meta-analysis of the Neurosynth database, using prior knowledge of regions in the temporal lobe defined by combining the NeuroFMA ontology with the Destrieux et al.'s atlas. We also produce a brain map for cognitive processes related to pain from a meta-analysis of the Neurosynth database, using prior knowledge of synonymous terms within 'biostimuli' subclasses defined in the IOBC ontology.
**Note**: the source code of NeuroLang will be open-sourced on GitHub by the time of the presentation.And here it is rendered
Presentors
Collaborators
Affiliations: @Parietal-INRIA, @neurospin
Abstract
Coordinate-based meta-analysis (CBMA) databases like Neurosynth are successfully used within brain mapping studies to define regions of interest supported by past literature. To the best of our knowledge, no tool is currently available to the brain mapping community for conveniently incorporating a priori neuroanatomical and ontological knowledge into richer meta-analyses. Through a language-oriented programming methodological approach, standing on decades of computer science research on logic programming languages and their probabilistic extensions, we build NeuroLang: a language for expressing rich multi-modal brain mapping hypotheses combining CBMA databases with neuroanatomical atlases (e.g. Destrieux et al.) and ontological knowledge bases (e.g. NeuroFMA).
In NeuroLang, CBMA databases, neuroanatomical atlases and ontological knowledge are seamlessly combined through declarative logic rules, forming a program. These rules are extended with probabilistic semantics to model the uncertainty inherent to noisy neuroimaging data. In all, NeuroLang is a fully-fledged probabilistic logic programming language. For instance, querying a NeuroLang program can produce forward inference maps that account for all the knowledge necessary to answer the query.
This demo walks you through examples of rich forward inference brain maps obtained from NeuroLang programs. We produce a brain map of auditory cognitive processes from a meta-analysis of the Neurosynth database, using prior knowledge of regions in the temporal lobe defined by combining the NeuroFMA ontology with the Destrieux et al.'s atlas. We also produce a brain map for cognitive processes related to pain from a meta-analysis of the Neurosynth database, using prior knowledge of synonymous terms within 'biostimuli' subclasses defined in the IOBC ontology.
Note: the source code of NeuroLang will be open-sourced on GitHub by the time of the presentation.
Thank you very much!
jsheunis
commented
4 years ago Done :)
choucavalier
commented
4 years ago This has been accepted as a Lightning! Thanks :) Happy to present
We will probably still make it more like a quick demo of our tool.
I confirm the title and abstracts are correct.
gzanitti
commented
4 years ago Slides: https://zenodo.org/record/3906015
For more information: https://neurolang.github.io/
Programming Brain Mapping Hypotheses in NeuroLang
By Valentin Iovene, Parietal, Inria Saclay / Neurospin, CEA
Abstract
Presentors
Collaborators
Affiliations: @Parietal-INRIA, @neurospin
Abstract
Coordinate-based meta-analysis (CBMA) databases like Neurosynth are successfully used within brain mapping studies to define regions of interest supported by past literature. To the best of our knowledge, no tool is currently available to the brain mapping community for conveniently incorporating a priori neuroanatomical and ontological knowledge into richer meta-analyses. Through a language-oriented programming methodological approach, standing on decades of computer science research on logic programming languages and their probabilistic extensions, we build NeuroLang: a language for expressing rich multi-modal brain mapping hypotheses combining CBMA databases with neuroanatomical atlases (e.g. Destrieux et al.) and ontological knowledge bases (e.g. NeuroFMA).
In NeuroLang, CBMA databases, neuroanatomical atlases and ontological knowledge are seamlessly combined through declarative logic rules, forming a program. These rules are extended with probabilistic semantics to model the uncertainty inherent to noisy neuroimaging data. In all, NeuroLang is a fully-fledged probabilistic logic programming language. For instance, querying a NeuroLang program can produce forward inference maps that account for all the knowledge necessary to answer the query.
This demo walks you through examples of rich forward inference brain maps obtained from NeuroLang programs. We produce a brain map of auditory cognitive processes from a meta-analysis of the Neurosynth database, using prior knowledge of regions in the temporal lobe defined by combining the NeuroFMA ontology with the Destrieux et al.'s atlas. We also produce a brain map for cognitive processes related to pain from a meta-analysis of the Neurosynth database, using prior knowledge of synonymous terms within 'biostimuli' subclasses defined in the IOBC ontology.
Note: the source code of NeuroLang will be open-sourced on GitHub by the time of the presentation.
Useful Links
https://github.com/gzanitti/Neurolang-OHBM2020-OSR
Tagging @tgy