Limits...
Using Graph Components Derived from an Associative Concept Dictionary to Predict fMRI Neural Activation Patterns that Represent the Meaning of Nouns.

Akama H, Miyake M, Jung J, Murphy B - PLoS ONE (2015)

Bottom Line: We measure the effectiveness of graph-based coefficients through the application of linguistic graph information for a neural activity recorded during conceptual processing in the human brain.Furthermore, correlating the voxel information with the MiF-based principal components, a new computational neurolinguistics model with a network connectivity paradigm is created.This allows two dimensions of context space to be incorporated with both semantic and neural distributional representations.

View Article: PubMed Central - PubMed

Affiliation: Graduate School of Decision Science and Technology, Tokyo Institute of Technology, Tokyo, Japan.

ABSTRACT
In this study, we introduce an original distance definition for graphs, called the Markov-inverse-F measure (MiF). This measure enables the integration of classical graph theory indices with new knowledge pertaining to structural feature extraction from semantic networks. MiF improves the conventional Jaccard and/or Simpson indices, and reconciles both the geodesic information (random walk) and co-occurrence adjustment (degree balance and distribution). We measure the effectiveness of graph-based coefficients through the application of linguistic graph information for a neural activity recorded during conceptual processing in the human brain. Specifically, the MiF distance is computed between each of the nouns used in a previous neural experiment and each of the in-between words in a subgraph derived from the Edinburgh Word Association Thesaurus of English. From the MiF-based information matrix, a machine learning model can accurately obtain a scalar parameter that specifies the degree to which each voxel in (the MRI image of) the brain is activated by each word or each principal component of the intermediate semantic features. Furthermore, correlating the voxel information with the MiF-based principal components, a new computational neurolinguistics model with a network connectivity paradigm is created. This allows two dimensions of context space to be incorporated with both semantic and neural distributional representations.

No MeSH data available.


Example of conceptual association overlaid on brain images representing its neural context.Top: lexical adjacency graph extracted from the semantic network of EAT (Fig 2). This represents MiF Principal Components (MiF-PCs) 3 (red labels) and 18 (black labels) with the fMRI nouns having the largest principal component scores (“bed” and “hand”, respectively) and the top twenty semantic features recording the largest principal component loading values. Most notably, the second fMRI noun for MiF-PC3 with the most sex-related connotation is also “hand”, so the graph shares various semantic contexts pertaining to this effector (body, sex, motions, and hand-carried goods). Bottom: anatomical location of the feature voxels selected from each participant of Mitchell et al. [26] as neural contexts corresponding to those two MiF-PCs. For example, “P1-PC3” denotes feature voxels from the P1 dataset that have neural activation patterns significantly homologous to the principal component vector of MiF-PC3 with respect to the 60 nouns used in the fMRI experiment. These sagittal brain images were smoothed using SPM8 with the full-width at half maximum parameter [333] to enhance visual effects. The Supporting Information and its figures clarify how to couple a neural component and an MiF-PC using an original fcMRI method applied to this semantico-neural paradigm.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4482269&req=5

pone.0125725.g007: Example of conceptual association overlaid on brain images representing its neural context.Top: lexical adjacency graph extracted from the semantic network of EAT (Fig 2). This represents MiF Principal Components (MiF-PCs) 3 (red labels) and 18 (black labels) with the fMRI nouns having the largest principal component scores (“bed” and “hand”, respectively) and the top twenty semantic features recording the largest principal component loading values. Most notably, the second fMRI noun for MiF-PC3 with the most sex-related connotation is also “hand”, so the graph shares various semantic contexts pertaining to this effector (body, sex, motions, and hand-carried goods). Bottom: anatomical location of the feature voxels selected from each participant of Mitchell et al. [26] as neural contexts corresponding to those two MiF-PCs. For example, “P1-PC3” denotes feature voxels from the P1 dataset that have neural activation patterns significantly homologous to the principal component vector of MiF-PC3 with respect to the 60 nouns used in the fMRI experiment. These sagittal brain images were smoothed using SPM8 with the full-width at half maximum parameter [333] to enhance visual effects. The Supporting Information and its figures clarify how to couple a neural component and an MiF-PC using an original fcMRI method applied to this semantico-neural paradigm.

Mentions: Fig 7 shows an example mapping for two circuits (or contexts) in parallel, i.e. conceptual relatedness with extending scope or growing complexity, and unexpectedly widespread fMRI responses to a lexical task, both associated with the nouns ‘bed’ and ‘hand’ (representative words for MiF-PC3: ‘bed’-HARD-SLEEVE-FINGER-SEX-LINING… and MiF-PC18: ‘hand’-CAP-BAG-SHOPPING BAG-WAVE-EXCHANGE…). Instead of determining some categorically-classified semantic atlas on the cortex (like “furniture” for ‘bed’ and “body parts” for ‘hand’), we generate a binding of informative voxels as a “neural context” (similar to a “semantic space” [71]), which serves as a counterpart to a lexical mapping of a key noun together with its semantic features. Note that all of these words are treated via fine-grained serial information as freely associated concepts under MiF-based principal components (extended to connotations such as sex, motions, and hand-carried goods; see the third column of S1 Table) that are intricate, context-sensitive, and in some way systematic.


Using Graph Components Derived from an Associative Concept Dictionary to Predict fMRI Neural Activation Patterns that Represent the Meaning of Nouns.

Akama H, Miyake M, Jung J, Murphy B - PLoS ONE (2015)

Example of conceptual association overlaid on brain images representing its neural context.Top: lexical adjacency graph extracted from the semantic network of EAT (Fig 2). This represents MiF Principal Components (MiF-PCs) 3 (red labels) and 18 (black labels) with the fMRI nouns having the largest principal component scores (“bed” and “hand”, respectively) and the top twenty semantic features recording the largest principal component loading values. Most notably, the second fMRI noun for MiF-PC3 with the most sex-related connotation is also “hand”, so the graph shares various semantic contexts pertaining to this effector (body, sex, motions, and hand-carried goods). Bottom: anatomical location of the feature voxels selected from each participant of Mitchell et al. [26] as neural contexts corresponding to those two MiF-PCs. For example, “P1-PC3” denotes feature voxels from the P1 dataset that have neural activation patterns significantly homologous to the principal component vector of MiF-PC3 with respect to the 60 nouns used in the fMRI experiment. These sagittal brain images were smoothed using SPM8 with the full-width at half maximum parameter [333] to enhance visual effects. The Supporting Information and its figures clarify how to couple a neural component and an MiF-PC using an original fcMRI method applied to this semantico-neural paradigm.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4482269&req=5

pone.0125725.g007: Example of conceptual association overlaid on brain images representing its neural context.Top: lexical adjacency graph extracted from the semantic network of EAT (Fig 2). This represents MiF Principal Components (MiF-PCs) 3 (red labels) and 18 (black labels) with the fMRI nouns having the largest principal component scores (“bed” and “hand”, respectively) and the top twenty semantic features recording the largest principal component loading values. Most notably, the second fMRI noun for MiF-PC3 with the most sex-related connotation is also “hand”, so the graph shares various semantic contexts pertaining to this effector (body, sex, motions, and hand-carried goods). Bottom: anatomical location of the feature voxels selected from each participant of Mitchell et al. [26] as neural contexts corresponding to those two MiF-PCs. For example, “P1-PC3” denotes feature voxels from the P1 dataset that have neural activation patterns significantly homologous to the principal component vector of MiF-PC3 with respect to the 60 nouns used in the fMRI experiment. These sagittal brain images were smoothed using SPM8 with the full-width at half maximum parameter [333] to enhance visual effects. The Supporting Information and its figures clarify how to couple a neural component and an MiF-PC using an original fcMRI method applied to this semantico-neural paradigm.
Mentions: Fig 7 shows an example mapping for two circuits (or contexts) in parallel, i.e. conceptual relatedness with extending scope or growing complexity, and unexpectedly widespread fMRI responses to a lexical task, both associated with the nouns ‘bed’ and ‘hand’ (representative words for MiF-PC3: ‘bed’-HARD-SLEEVE-FINGER-SEX-LINING… and MiF-PC18: ‘hand’-CAP-BAG-SHOPPING BAG-WAVE-EXCHANGE…). Instead of determining some categorically-classified semantic atlas on the cortex (like “furniture” for ‘bed’ and “body parts” for ‘hand’), we generate a binding of informative voxels as a “neural context” (similar to a “semantic space” [71]), which serves as a counterpart to a lexical mapping of a key noun together with its semantic features. Note that all of these words are treated via fine-grained serial information as freely associated concepts under MiF-based principal components (extended to connotations such as sex, motions, and hand-carried goods; see the third column of S1 Table) that are intricate, context-sensitive, and in some way systematic.

Bottom Line: We measure the effectiveness of graph-based coefficients through the application of linguistic graph information for a neural activity recorded during conceptual processing in the human brain.Furthermore, correlating the voxel information with the MiF-based principal components, a new computational neurolinguistics model with a network connectivity paradigm is created.This allows two dimensions of context space to be incorporated with both semantic and neural distributional representations.

View Article: PubMed Central - PubMed

Affiliation: Graduate School of Decision Science and Technology, Tokyo Institute of Technology, Tokyo, Japan.

ABSTRACT
In this study, we introduce an original distance definition for graphs, called the Markov-inverse-F measure (MiF). This measure enables the integration of classical graph theory indices with new knowledge pertaining to structural feature extraction from semantic networks. MiF improves the conventional Jaccard and/or Simpson indices, and reconciles both the geodesic information (random walk) and co-occurrence adjustment (degree balance and distribution). We measure the effectiveness of graph-based coefficients through the application of linguistic graph information for a neural activity recorded during conceptual processing in the human brain. Specifically, the MiF distance is computed between each of the nouns used in a previous neural experiment and each of the in-between words in a subgraph derived from the Edinburgh Word Association Thesaurus of English. From the MiF-based information matrix, a machine learning model can accurately obtain a scalar parameter that specifies the degree to which each voxel in (the MRI image of) the brain is activated by each word or each principal component of the intermediate semantic features. Furthermore, correlating the voxel information with the MiF-based principal components, a new computational neurolinguistics model with a network connectivity paradigm is created. This allows two dimensions of context space to be incorporated with both semantic and neural distributional representations.

No MeSH data available.