Class: Analysis¶
A distinct reported inferential map or tested effect, linking its source acquisitions, sample, paradigm context, tested effect, preprocessing, model estimation, inference settings, and source result tables.
Create one Analysis for each distinct tested effect. Separate Analyses are required when normalized direction, participant-group comparison, method, the pattern of effect.cells, seed or connectivity model, decoded variable, component identity, or spatial_scope differ, and when an omnibus effect and its directional post-hoc contrast are both reported. ROI definitions, acquisition summaries, descriptive sample tables, and masks without inferential effects are not Analyses. Nor is a map reported with no inferential test at all -- an ICA or PCA component map presented descriptively, a connectivity matrix given without a test -- there being nothing for such a map to be the effect of.
classDiagram
class Analysis
click Analysis href "../Analysis/"
Analysis : acquisitions
Analysis --> "*" Acquisition : acquisitions
click Acquisition href "../Acquisition/"
Analysis : assessments
Analysis --> "*" Assessment : assessments
click Assessment href "../Assessment/"
Analysis : coordinate_space
Analysis --> "0..1" ExtractedString : coordinate_space
click ExtractedString href "../ExtractedString/"
Analysis : defines_regions
Analysis --> "*" Region : defines_regions
click Region href "../Region/"
Analysis : definition
Analysis --> "1" ExtractedString : definition
click ExtractedString href "../ExtractedString/"
Analysis : details
Analysis --> "1" AnalysisDetails : details
click AnalysisDetails href "../AnalysisDetails/"
Analysis : effect
Analysis --> "1" Effect : effect
click Effect href "../Effect/"
Analysis : groups
Analysis --> "1..*" AnalysisGroup : groups
click AnalysisGroup href "../AnalysisGroup/"
Analysis : inference_settings
Analysis --> "0..1" InferenceSettings : inference_settings
click InferenceSettings href "../InferenceSettings/"
Analysis : interpretations
Analysis --> "0..1" ExtractedStringList : interpretations
click ExtractedStringList href "../ExtractedStringList/"
Analysis : local_id
Analysis : measure
Analysis --> "1" Measure : measure
click Measure href "../Measure/"
Analysis : mirror_of
Analysis --> "0..1" Analysis : mirror_of
click Analysis href "../Analysis/"
Analysis : model_estimation
Analysis --> "1" ModelEstimation : model_estimation
click ModelEstimation href "../ModelEstimation/"
Analysis : model_representation_notes
Analysis --> "0..1" ExtractedString : model_representation_notes
click ExtractedString href "../ExtractedString/"
Analysis : name
Analysis --> "1" ExtractedString : name
click ExtractedString href "../ExtractedString/"
Analysis : outcome
Analysis --> "0..1" ExtractedAnalysisOutcome : outcome
click ExtractedAnalysisOutcome href "../ExtractedAnalysisOutcome/"
Analysis : prespecification
Analysis --> "1" ExtractedPrespecification : prespecification
click ExtractedPrespecification href "../ExtractedPrespecification/"
Analysis : regions
Analysis --> "*" Region : regions
click Region href "../Region/"
Analysis : source_table_analysis
Analysis --> "0..1" ExtractedString : source_table_analysis
click ExtractedString href "../ExtractedString/"
Analysis : spatial_scope
Analysis --> "1" ExtractedSpatialScope : spatial_scope
click ExtractedSpatialScope href "../ExtractedSpatialScope/"
Analysis : stable_id
Analysis : tables
Analysis --> "*" Table : tables
click Table href "../Table/"
Analysis : tasks
Analysis --> "*" Task : tasks
click Task href "../Task/"
Slots¶
| Name | Cardinality and Range | Description | Inheritance |
|---|---|---|---|
| stable_id | 0..1 String |
UUID minted by the store the first time this entity appears, never by an extr... | direct |
| local_id | 1 String |
direct | |
| name | 1 ExtractedString |
Study-local label for the analysis or inferential result, in the source's wor... | direct |
| definition | 1 ExtractedString |
Precise statement of the tested effect in the source's own words, minimally c... | direct |
| prespecification | 1 ExtractedPrespecification |
Whether this contrast was planned before the data were seen or arrived at aft... | direct |
| interpretations | 0..1 ExtractedStringList |
The source's own statement of what this analysis found and what it means, typ... | direct |
| outcome | 0..1 ExtractedAnalysisOutcome |
Whether anything survived the threshold this analysis was tested at: a cluste... | direct |
| spatial_scope | 1 ExtractedSpatialScope |
Spatial domain of the analysis modelled | direct |
| regions | * Region |
The regions the analysis ran over -- its search space, what `spatial_scope: r... | direct |
| defines_regions | * Region |
Regions this analysis produced: a localizer's clusters, or peaks of a contras... | direct |
| measure | 1 Measure |
The scientific quantity of the map, by identifier, independent of the statist... | direct |
| groups | 1..* AnalysisGroup |
The participant cohorts the analysis ran on, and how many participants each c... | direct |
| effect | 1 Effect |
What was tested about the map | direct |
| details | 1 AnalysisDetails |
The method that produced the map and its method-specific payload | direct |
| acquisitions | * Acquisition |
Acquisition protocols supplying data to this analysis | direct |
| tasks | * Task |
The participant paradigms that contextualize this analysis | direct |
| model_estimation | 1 ModelEstimation |
The model this analysis's contrast was taken from: its family, estimator, sta... | direct |
| inference_settings | 0..1 InferenceSettings |
The thresholding and correction this analysis was reported under, by identifi... | direct |
| tables | * Table |
The source tables reporting results for this analysis | direct |
| source_table_analysis | 0..1 ExtractedString |
The key of the coordinate-parse analysis this analysis's results were read fr... | direct |
| coordinate_space | 0..1 ExtractedString |
Coordinate reference space or template the spatial results are reported in, s... | direct |
| assessments | * Assessment |
Assessments whose measurements were used in or selected the sample for this a... | direct |
| model_representation_notes | 0..1 ExtractedString |
Model components that could not be represented or were represented only appro... | direct |
| mirror_of | 0..1 Analysis |
Present only on an analysis the builder derived by reversing another one, and... | direct |
Usages¶
| used by | used in | type | used |
|---|---|---|---|
| Analysis | mirror_of | range | Analysis |
| CoordinateSet | analysis | range | Analysis |
| Study | analyses | range | Analysis |
Identifier and Mapping Information¶
Schema Source¶
- from schema: https://neurostuff.org/schemas/neuroimaging-study-extraction
Mappings¶
| Mapping Type | Mapped Value |
|---|---|
| self | neurostudy_ex:Analysis |
| native | neurostudy_ex:Analysis |
LinkML Source¶
Direct¶
name: Analysis
description: 'A distinct reported inferential map or tested effect, linking its source
acquisitions, sample, paradigm context, tested effect, preprocessing, model estimation,
inference settings, and source result tables.
Create one Analysis for each distinct tested effect. Separate Analyses are required
when normalized direction, participant-group comparison, method, the pattern of
`effect.cells`, seed or connectivity model, decoded variable, component identity,
or spatial_scope differ, and when an omnibus effect and its directional post-hoc
contrast are both reported. ROI definitions, acquisition summaries, descriptive
sample tables, and masks without inferential effects are not Analyses. Nor is a
map reported with no inferential test at all -- an ICA or PCA component map presented
descriptively, a connectivity matrix given without a test -- there being nothing
for such a map to be the effect of.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction
rank: 1000
attributes:
stable_id:
name: stable_id
description: UUID minted by the store the first time this entity appears, never
by an extractor, and kept across re-runs so a later extraction matches the same
entity.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
domain_of:
- Arm
- Timepoint
- Group
- Task
- Condition
- Assessment
- Region
- Acquisition
- Device
- Analysis
- Measure
- Preprocessing
- ModelEstimation
- ModelTerm
- InferenceSettings
- Table
- CoordinateSet
- Study
range: string
pattern: ^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$
local_id:
name: local_id
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
identifier: true
domain_of:
- Arm
- Timepoint
- Group
- Task
- Condition
- Assessment
- Region
- Acquisition
- Device
- Analysis
- Measure
- Preprocessing
- ModelEstimation
- ModelTerm
- InferenceSettings
- Table
- CoordinateSet
- Study
range: string
required: true
name:
name: name
description: Study-local label for the analysis or inferential result, in the
source's words.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
domain_of:
- Arm
- Timepoint
- Group
- Task
- Condition
- Assessment
- Region
- Analysis
- ModelTerm
- ExternalDataset
- PerformanceMetric
range: ExtractedString
required: true
definition:
name: definition
description: Precise statement of the tested effect in the source's own words,
minimally cleaned. Always source-grounded. For an interaction it must state
the direction of the interaction as the source states it; for an ordered contrast
it must keep the ordering that the cell directions cannot express.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
- ConjunctionComponent
range: ExtractedString
required: true
prespecification:
name: prespecification
description: 'Whether this contrast was planned before the data were seen or arrived
at afterwards.
Open vocabulary: write the source''s own wording in the field when neither value
fits. The two are not a spectrum and the words a source uses for them are many
-- `post-hoc`, `confirmatory`, `planned`, `hypothesis-driven` -- so a closed
range would discard the whole analysis over a synonym. Normalization maps the
wording onto the two values.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: ExtractedPrespecification
required: true
interpretations:
name: interpretations
description: 'The source''s own statement of what this analysis found and what
it means, typically in Results and Discussion, verbatim or minimally cleaned.
Always source-grounded, and the counterpart of `definition`: that field is the
Methods-side statement of what was tested, this one is the Results/Discussion-side
statement of what came of it.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: ExtractedStringList
outcome:
name: outcome
description: 'Whether anything survived the threshold this analysis was tested
at: a cluster, a peak, a significant region. The answer to "did this contrast
find anything", which `Effect.cells` cannot give -- a cell''s direction is the
sign of the contrast that was tested, and is written the same whether the test
found an effect or not. Meta-analyses exclude null effects and pool only analyses
that reported foci, so a tested-and-null contrast and a reported effect must
not extract to the same record. Applies only to an analysis whose coordinate
set is a `result` (or has no role yet): an ROI, seed, reference or other set
was not tested here, so it has no outcome.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: ExtractedAnalysisOutcome
spatial_scope:
name: spatial_scope
description: Spatial domain of the analysis modelled. The spatial_scope refers
to what voxels were modelled, not which voxels/brain regions had a correction
applied.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: ExtractedSpatialScope
required: true
regions:
name: regions
description: 'The regions the analysis ran over -- its search space, what `spatial_scope:
roi` restricts it to. Empty for whole-brain and searchlight, where the emptiness
asserts that inference was not restricted.
Not the analysis''s seeds: a whole-brain seed-based analysis has a seed and
no search-space region, so putting one here says the opposite of what it did.
Seeds are `ConnectivityDetails.seed_regions`.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
- FactorLevel
- CoordinateSet
- Study
range: Region
multivalued: true
inlined: false
defines_regions:
name: defines_regions
description: 'Regions this analysis produced: a localizer''s clusters, or peaks
of a contrast later used as seeds or ROIs. Populated on the analysis a region
came *from*, not on those that use it, so a region whose `definition_method`
is `same_study_analysis` and which another analysis lists in `regions` or `seed_regions`
is a circularity the record states rather than asserts.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: Region
multivalued: true
inlined: false
measure:
name: measure
description: 'The scientific quantity of the map, by identifier, independent of
the statistic used to test it. Analyses of one quantity reference one record;
an analysis of a different quantity takes its own. Required: a paper vague about
what it measured still takes a record, with `source_label` as given and `extraction_status:
not_reported` elsewhere.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: Measure
required: true
inlined: false
groups:
name: groups
description: 'The participant cohorts the analysis ran on, and how many participants
each contributed. Membership only: this slot says which people were analysed,
never what was compared. A cohort comparison is a pair of directional `Cell`s
on the model term whose levels name the cohorts, exactly as a condition or occasion
comparison is.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
- FactorLevel
- Study
range: AnalysisGroup
required: true
multivalued: true
inlined_as_list: true
effect:
name: effect
description: What was tested about the map. Universal across every method.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: Effect
required: true
inlined: true
details:
name: details
description: The method that produced the map and its method-specific payload.
The populated subclass identifies the method. A voxelwise, vertexwise, or ROI-wise
GLM, mixed-effects model, ANOVA, ANCOVA, or robust regression takes MassUnivariateDetails;
a method outside the first-class families takes OtherAnalysisDetails and names
itself in method_label.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: AnalysisDetails
required: true
inlined: true
acquisitions:
name: acquisitions
description: Acquisition protocols supplying data to this analysis. Structural,
diffusion, and other non-paradigm analyses are linked directly through this
field.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
domain_of:
- Task
- Analysis
- Study
range: Acquisition
multivalued: true
inlined: false
tasks:
name: tasks
description: 'The participant paradigms that contextualize this analysis. An analysis
can span paradigms, as a cross-task conjunction or a localizer-plus-main-task
analysis does.
A resting-state analysis names its Task here like any other: the resting-state
protocol is a paradigm, and a Task record exists for it. What leaves this empty
is an analysis with no paradigm at all -- morphometry, diffusion, a static PET
scan -- which is linked through `acquisitions` instead.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
- Study
range: Task
multivalued: true
inlined: false
model_estimation:
name: model_estimation
description: 'The model this analysis''s contrast was taken from: its family,
estimator, stage, software, and the terms it was specified with. What was selected
out of those terms is `effect`. Required: a paper that describes a contrast
and never names an estimator still takes a record, with the terms it does state
and `extraction_status: not_reported` elsewhere. Analyses sharing one design
reference one record; an analysis specified or estimated differently takes its
own.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: ModelEstimation
required: true
inlined: false
inference_settings:
name: inference_settings
description: The thresholding and correction this analysis was reported under,
by identifier. Analyses sharing one scheme reference one record; an analysis
thresholded or corrected differently takes its own. Not the model's -- one fit
thresholded two ways is two records here and one `model_estimation`. A contrast
reported at two search volumes is two analyses sharing one `model_estimation`,
not one analysis with the stricter threshold.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
- Study
range: InferenceSettings
inlined: false
tables:
name: tables
description: The source tables reporting results for this analysis.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
- Study
range: Table
multivalued: true
inlined: false
source_table_analysis:
name: source_table_analysis
description: 'The key of the coordinate-parse analysis this analysis''s results
were read from: `ParsedAnalysis.key` in neuroimaging-paper-parse, `analysis#<h>`,
where `h` hashes the parse analysis''s name, space and points. Derived from
what was extracted, not where it was read, so a re-parse that extracts the same
analysis reaches the same key. When pondie revised the parse, the key is the
revision''s.
This schema stores no coordinates, so reaching an analysis''s foci means joining
through the parse, and `tables` alone cannot do it: a table usually reports
several contrasts and several analyses usually cite the same table. This key
is what makes the join exact.
Copy it from the bracketed key the stage-1 listing prints beside each parsed
analysis. One key and not a list, because one analysis is read from one row
group; an analysis whose results span two tables cites both in `tables` and
names here the row group that defined it.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: ExtractedString
coordinate_space:
name: coordinate_space
description: 'Coordinate reference space or template the spatial results are reported
in, such as MNI152 or Talairach. **The authoritative of the two places this
can live**: where this and a `Table.coordinate_space` disagree, this one is
the analysis''s space, since a table can be parsed as one space while the analysis
it reports states another.
Both may be filled, and usually are: the table parser fills the tables and the
same parse supplies this when every table behind the analysis agrees, leaving
the model to answer only where they do not. Never inferred from the model''s
`spatial_unit` -- an analysis on a surface reports a surface template.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: ExtractedString
assessments:
name: assessments
description: Assessments whose measurements were used in or selected the sample
for this analysis.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
- Study
range: Assessment
multivalued: true
inlined: false
model_representation_notes:
name: model_representation_notes
description: Model components that could not be represented or were represented
only approximately, for an analysis whose method is otherwise first-class. Examples
include random slopes, latent variables, dynamic connectivity, and dependencies
between terms. An analysis whose method has no stable structured decomposition
at all takes NotStructurableDetails instead.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: ExtractedString
mirror_of:
name: mirror_of
description: 'Present only on an analysis the builder derived by reversing another
one, and naming the analysis it reverses.
A results table with both signs in one column reports one contrast read two
ways. The paper''s prose describes one direction, so only that half is extracted
and the other is rebuilt arithmetically -- same cells, directions flipped. This
slot is what says so. Without it the two halves look like independent findings,
and a meta-analysis that counts both counts one sample twice.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
domain_of:
- Analysis
range: Analysis
inlined: false
Induced¶
name: Analysis
description: 'A distinct reported inferential map or tested effect, linking its source
acquisitions, sample, paradigm context, tested effect, preprocessing, model estimation,
inference settings, and source result tables.
Create one Analysis for each distinct tested effect. Separate Analyses are required
when normalized direction, participant-group comparison, method, the pattern of
`effect.cells`, seed or connectivity model, decoded variable, component identity,
or spatial_scope differ, and when an omnibus effect and its directional post-hoc
contrast are both reported. ROI definitions, acquisition summaries, descriptive
sample tables, and masks without inferential effects are not Analyses. Nor is a
map reported with no inferential test at all -- an ICA or PCA component map presented
descriptively, a connectivity matrix given without a test -- there being nothing
for such a map to be the effect of.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction
rank: 1000
attributes:
stable_id:
name: stable_id
description: UUID minted by the store the first time this entity appears, never
by an extractor, and kept across re-runs so a later extraction matches the same
entity.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
owner: Analysis
domain_of:
- Arm
- Timepoint
- Group
- Task
- Condition
- Assessment
- Region
- Acquisition
- Device
- Analysis
- Measure
- Preprocessing
- ModelEstimation
- ModelTerm
- InferenceSettings
- Table
- CoordinateSet
- Study
range: string
pattern: ^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$
local_id:
name: local_id
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
identifier: true
owner: Analysis
domain_of:
- Arm
- Timepoint
- Group
- Task
- Condition
- Assessment
- Region
- Acquisition
- Device
- Analysis
- Measure
- Preprocessing
- ModelEstimation
- ModelTerm
- InferenceSettings
- Table
- CoordinateSet
- Study
range: string
required: true
name:
name: name
description: Study-local label for the analysis or inferential result, in the
source's words.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
owner: Analysis
domain_of:
- Arm
- Timepoint
- Group
- Task
- Condition
- Assessment
- Region
- Analysis
- ModelTerm
- ExternalDataset
- PerformanceMetric
range: ExtractedString
required: true
definition:
name: definition
description: Precise statement of the tested effect in the source's own words,
minimally cleaned. Always source-grounded. For an interaction it must state
the direction of the interaction as the source states it; for an ordered contrast
it must keep the ordering that the cell directions cannot express.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
- ConjunctionComponent
range: ExtractedString
required: true
prespecification:
name: prespecification
description: 'Whether this contrast was planned before the data were seen or arrived
at afterwards.
Open vocabulary: write the source''s own wording in the field when neither value
fits. The two are not a spectrum and the words a source uses for them are many
-- `post-hoc`, `confirmatory`, `planned`, `hypothesis-driven` -- so a closed
range would discard the whole analysis over a synonym. Normalization maps the
wording onto the two values.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: ExtractedPrespecification
required: true
interpretations:
name: interpretations
description: 'The source''s own statement of what this analysis found and what
it means, typically in Results and Discussion, verbatim or minimally cleaned.
Always source-grounded, and the counterpart of `definition`: that field is the
Methods-side statement of what was tested, this one is the Results/Discussion-side
statement of what came of it.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: ExtractedStringList
outcome:
name: outcome
description: 'Whether anything survived the threshold this analysis was tested
at: a cluster, a peak, a significant region. The answer to "did this contrast
find anything", which `Effect.cells` cannot give -- a cell''s direction is the
sign of the contrast that was tested, and is written the same whether the test
found an effect or not. Meta-analyses exclude null effects and pool only analyses
that reported foci, so a tested-and-null contrast and a reported effect must
not extract to the same record. Applies only to an analysis whose coordinate
set is a `result` (or has no role yet): an ROI, seed, reference or other set
was not tested here, so it has no outcome.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: ExtractedAnalysisOutcome
spatial_scope:
name: spatial_scope
description: Spatial domain of the analysis modelled. The spatial_scope refers
to what voxels were modelled, not which voxels/brain regions had a correction
applied.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: ExtractedSpatialScope
required: true
regions:
name: regions
description: 'The regions the analysis ran over -- its search space, what `spatial_scope:
roi` restricts it to. Empty for whole-brain and searchlight, where the emptiness
asserts that inference was not restricted.
Not the analysis''s seeds: a whole-brain seed-based analysis has a seed and
no search-space region, so putting one here says the opposite of what it did.
Seeds are `ConnectivityDetails.seed_regions`.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
- FactorLevel
- CoordinateSet
- Study
range: Region
multivalued: true
inlined: false
defines_regions:
name: defines_regions
description: 'Regions this analysis produced: a localizer''s clusters, or peaks
of a contrast later used as seeds or ROIs. Populated on the analysis a region
came *from*, not on those that use it, so a region whose `definition_method`
is `same_study_analysis` and which another analysis lists in `regions` or `seed_regions`
is a circularity the record states rather than asserts.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: Region
multivalued: true
inlined: false
measure:
name: measure
description: 'The scientific quantity of the map, by identifier, independent of
the statistic used to test it. Analyses of one quantity reference one record;
an analysis of a different quantity takes its own. Required: a paper vague about
what it measured still takes a record, with `source_label` as given and `extraction_status:
not_reported` elsewhere.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: Measure
required: true
inlined: false
groups:
name: groups
description: 'The participant cohorts the analysis ran on, and how many participants
each contributed. Membership only: this slot says which people were analysed,
never what was compared. A cohort comparison is a pair of directional `Cell`s
on the model term whose levels name the cohorts, exactly as a condition or occasion
comparison is.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
- FactorLevel
- Study
range: AnalysisGroup
required: true
multivalued: true
inlined: true
inlined_as_list: true
effect:
name: effect
description: What was tested about the map. Universal across every method.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: Effect
required: true
inlined: true
details:
name: details
description: The method that produced the map and its method-specific payload.
The populated subclass identifies the method. A voxelwise, vertexwise, or ROI-wise
GLM, mixed-effects model, ANOVA, ANCOVA, or robust regression takes MassUnivariateDetails;
a method outside the first-class families takes OtherAnalysisDetails and names
itself in method_label.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: AnalysisDetails
required: true
inlined: true
acquisitions:
name: acquisitions
description: Acquisition protocols supplying data to this analysis. Structural,
diffusion, and other non-paradigm analyses are linked directly through this
field.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
owner: Analysis
domain_of:
- Task
- Analysis
- Study
range: Acquisition
multivalued: true
inlined: false
tasks:
name: tasks
description: 'The participant paradigms that contextualize this analysis. An analysis
can span paradigms, as a cross-task conjunction or a localizer-plus-main-task
analysis does.
A resting-state analysis names its Task here like any other: the resting-state
protocol is a paradigm, and a Task record exists for it. What leaves this empty
is an analysis with no paradigm at all -- morphometry, diffusion, a static PET
scan -- which is linked through `acquisitions` instead.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
- Study
range: Task
multivalued: true
inlined: false
model_estimation:
name: model_estimation
description: 'The model this analysis''s contrast was taken from: its family,
estimator, stage, software, and the terms it was specified with. What was selected
out of those terms is `effect`. Required: a paper that describes a contrast
and never names an estimator still takes a record, with the terms it does state
and `extraction_status: not_reported` elsewhere. Analyses sharing one design
reference one record; an analysis specified or estimated differently takes its
own.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: ModelEstimation
required: true
inlined: false
inference_settings:
name: inference_settings
description: The thresholding and correction this analysis was reported under,
by identifier. Analyses sharing one scheme reference one record; an analysis
thresholded or corrected differently takes its own. Not the model's -- one fit
thresholded two ways is two records here and one `model_estimation`. A contrast
reported at two search volumes is two analyses sharing one `model_estimation`,
not one analysis with the stricter threshold.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
- Study
range: InferenceSettings
inlined: false
tables:
name: tables
description: The source tables reporting results for this analysis.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
- Study
range: Table
multivalued: true
inlined: false
source_table_analysis:
name: source_table_analysis
description: 'The key of the coordinate-parse analysis this analysis''s results
were read from: `ParsedAnalysis.key` in neuroimaging-paper-parse, `analysis#<h>`,
where `h` hashes the parse analysis''s name, space and points. Derived from
what was extracted, not where it was read, so a re-parse that extracts the same
analysis reaches the same key. When pondie revised the parse, the key is the
revision''s.
This schema stores no coordinates, so reaching an analysis''s foci means joining
through the parse, and `tables` alone cannot do it: a table usually reports
several contrasts and several analyses usually cite the same table. This key
is what makes the join exact.
Copy it from the bracketed key the stage-1 listing prints beside each parsed
analysis. One key and not a list, because one analysis is read from one row
group; an analysis whose results span two tables cites both in `tables` and
names here the row group that defined it.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: ExtractedString
coordinate_space:
name: coordinate_space
description: 'Coordinate reference space or template the spatial results are reported
in, such as MNI152 or Talairach. **The authoritative of the two places this
can live**: where this and a `Table.coordinate_space` disagree, this one is
the analysis''s space, since a table can be parsed as one space while the analysis
it reports states another.
Both may be filled, and usually are: the table parser fills the tables and the
same parse supplies this when every table behind the analysis agrees, leaving
the model to answer only where they do not. Never inferred from the model''s
`spatial_unit` -- an analysis on a surface reports a surface template.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: ExtractedString
assessments:
name: assessments
description: Assessments whose measurements were used in or selected the sample
for this analysis.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
- Study
range: Assessment
multivalued: true
inlined: false
model_representation_notes:
name: model_representation_notes
description: Model components that could not be represented or were represented
only approximately, for an analysis whose method is otherwise first-class. Examples
include random slopes, latent variables, dynamic connectivity, and dependencies
between terms. An analysis whose method has no stable structured decomposition
at all takes NotStructurableDetails instead.
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: ExtractedString
mirror_of:
name: mirror_of
description: 'Present only on an analysis the builder derived by reversing another
one, and naming the analysis it reverses.
A results table with both signs in one column reports one contrast read two
ways. The paper''s prose describes one direction, so only that half is extracted
and the other is rebuilt arithmetically -- same cells, directions flipped. This
slot is what says so. Without it the two halves look like independent findings,
and a meta-analysis that counts both counts one sample twice.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction/analysis
rank: 1000
owner: Analysis
domain_of:
- Analysis
range: Analysis
inlined: false