Package 'PhysioMSKNet'

Description

Computes the degree distribution for bones or muscles.

Usage

degreeDistribution(hg, type = c("muscle", "bone"))

Arguments

Value

A data.frame with columns: degree, count, probability

Description

Detects functional communities from EMG coherence and compares them with structural communities from the MSK network.

Usage

emgCommunityCompare(pe, hg = NULL, gamma = 4.3, mapping = NULL)

Arguments

Value

A list with:

emg_communities

Named integer vector of EMG community assignments

msk_communities

Named integer vector of MSK community assignments

z_rand

z-Rand score comparing the two partitions

mapping

The channel-to-muscle mapping used

Description

Tests whether EMG activation levels differ across MSK community assignments using a Kruskal-Wallis test.

Usage

emgMSKEnrichment(pe, hg = NULL, gamma = 4.3, mapping = NULL)

Arguments

Value

A list with:

per_community

Data frame with community, mean/median activation

overall_test

Kruskal-Wallis test result

activation_values

Named numeric vector of RMS activation

Description

Computes EMG functional coherence and compares it with structural adjacency from the MSK muscle graph using a Mantel test.

Usage

emgStructuralCoherence(pe, hg = NULL, freq_band = c(20, 50), mapping = NULL)

Arguments

Value

A list with:

coherence_matrix

EMG functional coherence matrix

structural_matrix

MSK structural adjacency (matched subset)

correlation

Mantel correlation coefficient

p_value

Permutation-based p-value

mapped_muscles

Names of matched muscles

Examples

## Not run: 
result <- emgStructuralCoherence(pe_emg, freq_band = c(20, 50))

## End(Not run)

Description

Matches EMG channel names from a PhysioExperiment/SummarizedExperiment object to muscles in an MSK hypergraph.

Usage

emgToMSKMapping(pe, hg = NULL, method = c("exact", "fuzzy"), threshold = 0.8)

Arguments

Value

A data.frame with columns: channel_idx, channel_name, muscle_idx, muscle_name, match_quality.

Examples

## Not run: 
mapping <- emgToMSKMapping(pe_emg, method = "fuzzy")

## End(Not run)

Description

Returns the degree of each hyperedge (muscle), i.e., the number of bones each muscle attaches to.

Usage

hyperedgeDegree(hg)

Arguments

Value

Named numeric vector of muscle degrees.

Description

Analyzes movement synchrony from IMU sensors in the context of MSK network communities. Computes within-community vs between-community sensor synchrony and tests for significance.

Usage

imuCommunityDynamics(
  imu_data,
  hg = NULL,
  gamma = 4.3,
  mapping = NULL,
  window_sec = NULL
)

Arguments

Value

A list with:

within_community_sync

Mean synchrony within communities

between_community_sync

Mean synchrony between communities

ratio

Within/between ratio

p_value

Wilcoxon test p-value

time_resolved

Data frame if window_sec provided, NULL otherwise

Description

Computes bone stress from IMU acceleration/angular velocity data and propagates it through the MSK incidence matrix to predict muscle vulnerability.

Usage

imuImpactPrediction(
  imu_data,
  hg = NULL,
  mapping = NULL,
  stress_metric = c("acceleration", "angular_velocity", "jerk", "composite"),
  use_proxy = TRUE
)

Arguments

Value

A list with:

vulnerability

Named numeric vector of muscle vulnerability scores

bone_stress

Named numeric vector of bone stress

muscle_stress_exposure

Named numeric of stress propagated to muscles

ranking

Data frame ranked by vulnerability (descending)

Description

Computes pairwise kinematic coupling from IMU orientation or acceleration data and compares it to the structural adjacency of the MSK bone graph using a Mantel test.

Usage

imuNetworkKinematics(
  imu_data,
  hg = NULL,
  mapping = NULL,
  signal = c("orientation", "acceleration"),
  method = c("correlation", "mutual_info")
)

Arguments

Value

A list with:

kinematic_coupling

Pairwise coupling matrix between mapped bones

structural_matrix

Corresponding MSK bone graph adjacency

correlation

Mantel test correlation coefficient

p_value

Permutation p-value

mapped_sensors

Data frame of sensor-to-bone mapping used

Description

Matches IMU sensor placement names to bones in the MSK hypergraph using a curated lookup table with fuzzy matching fallback. Supports naming conventions from Xsens, APDM, Shimmer, and generic body-segment labels.

Usage

imuToMSKMapping(
  sensors,
  hg = NULL,
  method = c("exact", "fuzzy"),
  threshold = 0.7
)

Arguments

Value

A data.frame with columns: sensor_name, bone_idx, bone_name, match_quality, match_method.

Examples

## Not run: 
mapping <- imuToMSKMapping(c("upper_arm", "thigh", "lumbar"))

## End(Not run)

Description

Loads the bipartite incidence matrix C where rows are bones (vertices) and columns are muscles (hyperedges). Entry C[i,j] = 1 indicates that muscle j attaches to bone i.

Usage

loadIncidenceMatrix()

Value

A sparse Matrix (dgCMatrix) of dimensions 173 x 270

References

Murphy AC et al. (2018) PLOS Biology.

Description

Loads the complete musculoskeletal network dataset from Murphy et al. (2018). Returns a list containing the incidence matrix, muscle metadata, and validation data for reproducing the paper's results.

Usage

loadMSKData()

Value

A list with components:

incidence

Sparse incidence matrix C (bones x muscles)

muscle_meta

Data frame with muscle names, community assignments, homunculus categories

bone_names

Character vector of bone/vertex names

muscle_names

Character vector of muscle/hyperedge names

References

Murphy AC et al. (2018) "Structure, function, and control of the human musculoskeletal network." PLOS Biology 16(1): e2002811.

Description

Loads metadata for all 270 muscles including community assignments and homunculus category labels from the paper.

Usage

loadMuscleMetadata()

Value

A data.frame with columns: index, muscle, community, homunculus_category

References

Murphy AC et al. (2018) PLOS Biology.

Description

Loads validation datasets used to reproduce key figures from the paper.

Usage

loadValidationData(
  dataset = c("degree_distribution", "impact_vs_recovery", "homunculus_deviation",
    "fmri_activation", "homunculus_coordinates", "impact_vs_path", "impact_scores")
)

Arguments

Value

A data.frame with the requested validation data

References

Murphy AC et al. (2018) PLOS Biology.

Description

Computes displacement synchrony (correlation) between matched bones and tests whether within-community pairs are more synchronized than between-community pairs using a Wilcoxon test.

Usage

mocapCommunityDynamics(
  pe_mocap,
  hg = NULL,
  gamma = 4.3,
  mapping = NULL,
  window_sec = NULL
)

Arguments

Value

A list with:

within_community_sync

Mean within-community synchrony

between_community_sync

Mean between-community synchrony

ratio

Within/between ratio

p_value

Wilcoxon test p-value

time_resolved

Data frame with per-window results (if window_sec set)

Description

Computes kinematic stress per bone from MoCap data and propagates it through the MSK incidence matrix to estimate muscle vulnerability.

Usage

mocapImpactPrediction(
  pe_mocap,
  hg = NULL,
  mapping = NULL,
  stress_metric = c("acceleration", "jerk", "range"),
  use_proxy = TRUE
)

Arguments

Value

A list with:

vulnerability

Named numeric vector of muscle vulnerability scores

bone_stress

Named numeric vector of bone stress values

muscle_stress_exposure

Named numeric vector of muscle stress exposure

ranking

Data frame ranking muscles by vulnerability

Description

Computes pairwise movement coupling between MoCap segments and compares the kinematic coupling matrix with MSK structural adjacency.

Usage

mocapNetworkKinematics(
  pe_mocap,
  hg = NULL,
  mapping = NULL,
  method = c("correlation", "mutual_info")
)

Arguments

Value

A list with:

kinematic_coupling

Pairwise coupling matrix

structural_matrix

MSK bone adjacency (matched subset)

correlation

Mantel correlation coefficient

p_value

Permutation-based p-value

Description

Matches MoCap segment names from a SkeletonModel or character vector to bones in an MSK hypergraph using a curated lookup table with fuzzy matching fallback.

Usage

mocapToMSKMapping(
  skeleton,
  hg = NULL,
  method = c("exact", "fuzzy"),
  threshold = 0.7
)

Arguments

Value

A data.frame with columns: segment_name, bone_idx, bone_name, match_quality, match_method.

Examples

## Not run: 
mapping <- mocapToMSKMapping(c("upper_arm", "forearm", "thigh"))

## End(Not run)

Description

Modifies a rehabilitation protocol based on patient reassessment results. Uses decision rules to determine whether to advance, continue, reduce, or modify the current protocol.

Usage

mskAdaptProtocol(reassessment, current_protocol, hg = NULL)

Arguments

Value

An S3 object of class "MSKAdaptedProtocol" with:

decision

character: "advance", "continue", "reduce", "modify"

adapted_exercises

data.frame with modified exercise prescription

rationale

character explanation

focus_muscles

character vector of muscles needing attention

removed_exercises

exercises to drop

added_exercises

new exercises to add

Examples

## Not run: 
protocol <- mskRehabProtocol("Biceps Brachii")
adapted <- mskAdaptProtocol(reassessment, protocol)
print(adapted)

## End(Not run)

Description

Attaches anatomical annotations from PhysioAnnotationHub to an MSK hypergraph, matching muscle and bone names between the two data sources. Annotations include body region, innervation, actions, and spinal levels.

Usage

mskAnnotate(hg = NULL, hub = NULL)

Arguments

Value

An annotated MSKHypergraph with additional fields:

muscle_annotations

Data frame of muscle annotations merged from hub

bone_annotations

Data frame of bone annotations merged from hub

annotated

Logical flag indicating annotations are attached

annotation_coverage

List with muscle and bone match rates

Clinical Validity

Annotations are derived from PhysioAnnotationHub's curated knowledge graph. Name matching uses fuzzy matching with a 0.2 edit distance threshold, which may produce incorrect matches for similarly named structures. Always verify critical annotations against primary anatomical references.

References

Murphy AC et al. (2018) PLOS Biology 16(1): e2002811.

Examples

## Not run: 
hg <- mskAnnotate()
head(hg$muscle_annotations)

## End(Not run)

Description

Computes betweenness centrality for the projected graph. Uses igraph if available, otherwise a pure-R BFS implementation.

Usage

mskBetweenness(hg, type = c("bone", "muscle"))

Arguments

Value

Named numeric vector of betweenness centrality values.

Description

Generates a clinical evidence report for a set of injured muscles by querying the knowledge graph for ICD-10 codes, ICF codes, innervation, functional impact, and related muscles (synergists and antagonists).

Usage

mskClinicalEvidence(injury_muscles, hub = NULL, hg = NULL)

Arguments

Value

An S3 object of class "MSKClinicalEvidence" with:

muscles

Data frame of injury muscle annotations

icd10_codes

Matching ICD-10 entries

icf_codes

Matching ICF entries

affected_nerves

Nerves innervating injured muscles

affected_spinal_levels

Spinal segments involved

functional_impact

Affected actions/movements

synergists

Synergistic muscles from KG

antagonists

Antagonistic muscles from KG

Clinical Validity

ICD-10 and ICF code mappings are based on knowledge graph associations and may not capture all valid codes for a clinical scenario. This tool provides evidence aggregation for research; clinical coding should follow local guidelines and be reviewed by qualified professionals.

Examples

## Not run: 
evidence <- mskClinicalEvidence(c("Biceps Brachii", "Deltoid"))
print(evidence)

## End(Not run)

Description

Predicts recovery time, identifies compensatory muscles, and estimates secondary injury risk based on MSK network impact analysis.

Usage

mskClinicalPredictor(injury_muscles, hg = NULL, sim = NULL, verbose = TRUE)

Arguments

Value

An S3 object of class "MSKClinicalPrediction" with:

recovery

Data frame with predicted recovery weeks and CI per muscle

compensatory

List of compensatory muscles per injured muscle

secondary_risk

Data frame of secondary injury risk scores

injury_muscles

Resolved muscle names

injury_indices

Resolved integer indices

Clinical Validity

The recovery model was validated on 14 aggregate muscle groups, not individual muscles. Patient factor adjustments are heuristic, not independently validated. This is a research exploration tool, not a clinical diagnostic.

References

Murphy AC et al. (2018) PLOS Biology 16(1): e2002811.

Examples

## Not run: 
pred <- mskClinicalPredictor(c("Biceps Brachii", "Deltoid"))
print(pred)

## End(Not run)

Description

Closeness Centrality

Usage

mskCloseness(hg, type = c("bone", "muscle"))

Arguments

Value

Named numeric vector of closeness centrality values.

Description

Detects communities in the projected muscle-muscle graph using the Louvain algorithm with a resolution parameter gamma.

Usage

mskCommunityDetect(hg = NULL, gamma = 4.3, type = c("muscle", "bone"))

Arguments

Value

A list with:

membership

Named integer vector of community assignments

n_communities

Number of communities detected

modularity

Modularity value Q

gamma

Resolution parameter used

sizes

Table of community sizes

References

Murphy AC et al. (2018) PLOS Biology 16(1): e2002811.

Description

Generates a detailed functional profile of a musculoskeletal community by combining network topology metrics with knowledge graph annotations. This enables characterization of communities by their anatomical region, innervation patterns, primary actions, and spinal segment involvement.

Usage

mskCommunityProfile(hg = NULL, community_id, hub = NULL, gamma = 4.3)

Arguments

Value

An S3 object of class "MSKCommunityProfile" with:

community_id

Integer, the profiled community

muscles

Character vector of muscle names in the community

n_muscles

Integer, number of muscles

action_profile

Table of primary action distribution

nerve_profile

Table of innervating nerves distribution

region_profile

Table of body region distribution

spinal_profile

Table of spinal level distribution

dominant_action

Most frequent primary action

dominant_nerve

Most frequent innervating nerve

dominant_region

Most frequent body region

mean_degree

Mean hyperedge degree of community muscles

mean_impact_deviation

Mean impact deviation score

Clinical Validity

Community membership depends on the resolution parameter gamma and the stochastic Louvain algorithm. Profile annotations depend on KG completeness. Community boundaries are network-derived, not anatomical boundaries.

References

Murphy AC et al. (2018) PLOS Biology 16(1): e2002811.

Examples

## Not run: 
profile <- mskCommunityProfile(community_id = 1)
print(profile)

## End(Not run)

Description

Tracks changes in compensation patterns across multiple timepoints, identifying onset, resolution, and trends.

Usage

mskCompensationEvolution(
  timepoints_emg,
  injured_muscles,
  hg = NULL,
  emg_mapping = NULL,
  sr = NULL,
  z_threshold = 1.96
)

Arguments

Value

An S3 object of class "MSKCompensationEvolution" with:

evolution_table

Data.frame of per-timepoint per-muscle z-scores

onset_timepoint

Per-muscle first timepoint of compensation

resolution_timepoint

Per-muscle first timepoint of resolution

trend

Per-muscle trend classification

summary

Data.frame of per-timepoint summary statistics

Examples

## Not run: 
evolution <- mskCompensationEvolution(
  timepoints_emg = list(week0 = emg0, week2 = emg2, week4 = emg4),
  injured_muscles = c("Biceps Brachii")
)

## End(Not run)

Description

Constructs a compensation-weighted subgraph from compensation detection results, identifying compensation chains and hub muscles.

Usage

mskCompensationNetwork(compensation_result, hg = NULL, emg_mapping = NULL)

Arguments

Value

A list with:

adjacency

Weighted compensation adjacency matrix

chains

List of compensation chains

hub_muscles

Muscles in multiple compensation chains

community_involvement

MSK communities affected by compensation

Examples

## Not run: 
net <- mskCompensationNetwork(compensation_result, hg)

## End(Not run)

Description

Computes risk scores for compensating muscles based on biomechanical overuse principles, integrating network topology, duration of compensation, and loading intensity.

Usage

mskCompensationRiskScore(
  compensation_result,
  hg = NULL,
  duration_weeks = 0,
  load_intensity = c("low", "moderate", "high")
)

Arguments

Value

An S3 object of class "MSKCompensationRisk" with:

per_muscle_risk

Data.frame of per-muscle risk scores

overall_risk

Numeric overall risk score

overall_category

Character risk category

highest_risk_muscle

Name of highest-risk muscle

recommendation

Clinical action recommendation

Risk Model

Per compensating muscle: risk = z_excess * degree_normalized * duration_factor * load_factor where z_excess = max(0, z_score - z_threshold), degree_normalized = hyperedgeDegree / mean_degree, duration_factor = 1 + log(1 + duration_weeks), load_factor = intensity multiplier (0.5, 1.0, 1.5).

Examples

## Not run: 
risk <- mskCompensationRiskScore(compensation_result,
  duration_weeks = 4, load_intensity = "moderate")

## End(Not run)

Description

Orchestrator that runs all compensation analyses and returns a unified summary. Uses tryCatch for each sub-analysis so partial results are available even if some analyses fail.

Usage

mskCompensationSummary(
  emg,
  emg_baseline,
  injured_muscles,
  timepoints_emg = NULL,
  hg = NULL,
  sr = NULL,
  z_threshold = 1.96,
  duration_weeks = 0,
  load_intensity = c("low", "moderate", "high")
)

Arguments

Value

An S3 object of class "MSKCompensationSummary" with sub-results and available_analyses vector.

Examples

## Not run: 
summary <- mskCompensationSummary(
  emg = emg_current, emg_baseline = emg_pre,
  injured_muscles = c("Biceps Brachii"),
  duration_weeks = 4, load_intensity = "moderate"
)

## End(Not run)

Description

Runs Louvain community detection multiple times and extracts a consensus partition. The paper uses 100 runs and takes the most frequent partition for each node pair.

Usage

mskConsensusPartition(
  hg = NULL,
  gamma = 4.3,
  n_runs = 100L,
  type = c("muscle", "bone")
)

Arguments

Value

Same structure as mskCommunityDetect, but with consensus partition.

References

Murphy AC et al. (2018) PLOS Biology.

Description

Compares a subject's EMG coordination pattern against a reference (healthy baseline or normative data) to produce a 0-1 quality score.

Usage

mskCoordinationQualityScore(
  emg,
  reference_emg = NULL,
  hg = NULL,
  method = c("synergy_distance", "correlation_profile", "network_similarity")
)

Arguments

Value

A list with: quality_score (0-1, 1=perfect match to reference), component_scores, muscle_contributions.

Examples

## Not run: 
cqs <- mskCoordinationQualityScore(emg, reference_emg, method = "synergy_distance")

## End(Not run)

Description

Detects compensatory movement patterns by comparing current EMG activation against a baseline, using MSK network topology to identify muscles that are structurally positioned to compensate for injured muscles.

Usage

mskDetectCompensation(
  emg,
  emg_baseline,
  injured_muscles,
  hg = NULL,
  emg_mapping = NULL,
  sr = NULL,
  z_threshold = 1.96,
  neighborhood_order = 2L
)

Arguments

Value

An S3 object of class "MSKCompensation" with:

compensating_muscles

Data.frame of muscles showing compensatory activation

injured_status

Data.frame of injured muscle activation status

non_compensating

Data.frame of neighbors that did NOT compensate

compensation_prevalence

Proportion of neighbors showing compensation

network_context

List with neighborhood info

Algorithm

1. Map EMG channels to MSK muscles via emgToMSKMapping

2. Compute RMS activation for both current and baseline

3. Compute z-score of change: z = (current_rms - baseline_rms) / baseline_sd

4. Identify MSK neighbors of injured muscles within neighborhood_order hops

5. Flag muscles where they are neighbors AND z-score > z_threshold

6. Also detect decreased activation in injured muscles (z < -z_threshold)

Examples

## Not run: 
result <- mskDetectCompensation(
  emg = emg_current, emg_baseline = emg_pre,
  injured_muscles = c("Biceps Brachii"),
  z_threshold = 1.96
)
print(result)

## End(Not run)

Description

Identifies when recovery has plateaued using rolling window slope analysis or change rate assessment.

Usage

mskDetectRecoveryPlateau(
  tracker,
  window = 3L,
  min_slope = NULL,
  method = c("slope", "change_rate")
)

Arguments

Value

An S3 object of class "MSKRecoveryPlateau" with:

plateau_detected

logical per muscle

plateau_onset

timepoint index where plateau begins (NA if none)

details

data.frame with per-window slopes

recommendation

character ("continue", "modify_protocol", "reassess")

Examples

## Not run: 
plateau <- mskDetectRecoveryPlateau(tracker, window = 3)

## End(Not run)

Description

Compares first and last timepoint to determine if each muscle shows clinically meaningful change beyond the MDC threshold.

Usage

mskDetectResponderStatus(
  tracker,
  mdc_result = NULL,
  threshold_type = c("mdc", "effect_size")
)

Arguments

Value

An S3 object of class "MSKResponderStatus" with:

classification

data.frame (muscle, status, change, threshold, effect_size_d)

summary

counts of responders/non-responders/deteriorated

overall_status

"responder" if majority of muscles improve

Examples

## Not run: 
status <- mskDetectResponderStatus(tracker, mdc_result)

## End(Not run)

Description

Tests whether a set of muscles is enriched for specific anatomical or functional annotations (actions, nerves, body regions, spinal levels) compared to the full background set of muscles in the hypergraph.

Usage

mskEnrichKG(
  muscles,
  annotation_type = c("action", "nerve", "body_region", "spinal_level"),
  hub = NULL,
  hg = NULL
)

Arguments

Value

A data.frame with columns:

term

Annotation term

count

Number of query muscles with this term

background_count

Number of background muscles with this term

expected

Expected count under null

fold_enrichment

Ratio of observed to expected

p_value

P-value from hypergeometric test

significant

Logical, TRUE if p < 0.05

Clinical Validity

Enrichment analysis depends on the completeness and accuracy of the underlying knowledge graph annotations. P-values are not adjusted for multiple testing; consider applying Bonferroni or FDR correction when testing multiple annotation types simultaneously.

Examples

## Not run: 
# Test enrichment of upper limb muscles for nerve innervation
enrichment <- mskEnrichKG(
  c("Biceps Brachii", "Deltoid", "Trapezius"),
  annotation_type = "nerve"
)
enrichment[enrichment$significant, ]

## End(Not run)

Description

Takes the output of mskPredictFunctionalOutcome and generates time-based milestones for tracking rehabilitation progress.

Usage

mskFunctionalMilestones(
  outcome_prediction,
  n_milestones = 4L,
  milestone_type = c("clinical", "linear", "accelerating")
)

Arguments

Value

An S3 object of class "MSKFunctionalMilestones" with:

milestones

data.frame with milestone_id, week, target_rom, target_strength, target_function, phase_name, description

timeline_weeks

total timeline in weeks

milestone_type

character

Examples

## Not run: 
outcome <- mskPredictFunctionalOutcome("Biceps Brachii")
milestones <- mskFunctionalMilestones(outcome, milestone_type = "clinical")
print(milestones)

## End(Not run)

Description

Tests the correspondence between MSK network community structure and the motor cortex homunculus. Reproduces Fig 4b results. Target: F(1,19) = 21.3, R² = 0.52, p < 0.001

Usage

mskHomuncCorrelation(membership = NULL, homunculus_data = NULL)

Arguments

Value

A list with regression result and deviation ratio by category.

References

Murphy AC et al. (2018) PLOS Biology.

Description

Constructs a musculoskeletal hypergraph from an incidence matrix. In this hypergraph, bones are vertices and muscles are hyperedges. A muscle (hyperedge) connects all bones it attaches to.

Usage

MSKHypergraph(C = NULL, muscle_meta = NULL)

Arguments

Value

An S3 object of class "MSKHypergraph" containing:

C

Sparse incidence matrix (bones x muscles)

n_bones

Number of bones (vertices)

n_muscles

Number of muscles (hyperedges)

bone_names

Character vector of bone names

muscle_names

Character vector of muscle names

muscle_meta

Muscle metadata if provided

References

Murphy AC et al. (2018) PLOS Biology 16(1): e2002811.

Description

Computes the impact deviation for each muscle, which is the difference between the observed impact score and the expected impact score for a muscle of that degree, expressed in standard deviations.

Usage

mskImpactDeviation(impact_scores, hg, null_scores = NULL)

Arguments

Details

The expected impact score is estimated from null model simulations or from a regression of impact score vs. degree.

Value

A named numeric vector of impact deviations.

References

Murphy AC et al. (2018) PLOS Biology.

Description

Reproduces the key result from Fig 3b: correlation between impact deviation and clinical muscle injury recovery time. Target: F(1,12) = 37.3, R² = 0.757, p < 0.0001

Usage

mskImpactRecoveryModel(impact_deviation = NULL, recovery_data = NULL)

Arguments

Value

Result from mskRobustRegression with additional paper comparison.

References

Murphy AC et al. (2018) PLOS Biology Table 4.

Description

Perturbs a single muscle and simulates the network dynamics to compute the total displacement of all bones (impact score). The perturbation is applied in the 4th spatial dimension to avoid directional artifacts.

Usage

mskImpactScore(sim, muscle_index)

Arguments

Value

Numeric impact score (total displacement summed over all bones).

References

Murphy AC et al. (2018) PLOS Biology.

Description

Runs perturbation analysis for all 270 muscles and returns impact scores. This is the main computational function for reproducing Fig. 3a of the paper.

Usage

mskImpactScoreAll(sim = NULL, verbose = TRUE)

Arguments

Value

A named numeric vector of impact scores for each muscle.

References

Murphy AC et al. (2018) PLOS Biology.

Description

Computes a personalized injury risk profile for all muscles based on MSK network topology and patient characteristics.

Usage

mskInjuryRiskProfile(patient_data, hg = NULL, sim = NULL)

Arguments

Value

An S3 object of class "MSKInjuryRiskProfile" with:

risk_scores

Data frame with muscle name, base risk, adjusted risk

patient_data

Input patient data

factors

Applied adjustment factors

Clinical Validity

Patient factor adjustments (age, BMI, activity level) are heuristic multipliers, not derived from validated epidemiological models. This is a research exploration tool, not a clinical diagnostic.

Examples

## Not run: 
profile <- mskInjuryRiskProfile(list(age = 45, activity_level = "active"))

## End(Not run)

Description

Produces a comprehensive summary of the MSK network annotated with knowledge graph data. Profiles all communities, identifies cross-community nerve pathways, and summarizes annotation coverage.

Usage

mskKGSummary(hg = NULL, hub = NULL, gamma = 4.3)

Arguments

Value

An S3 object of class "MSKKGSummary" with:

hg

Annotated MSKHypergraph

community_profiles

List of MSKCommunityProfile objects

n_communities

Number of detected communities

cross_community_nerves

Data frame of nerves spanning communities

annotation_coverage

Annotation match statistics

Clinical Validity

This is a summary tool combining network topology with knowledge graph annotations. Community boundaries and annotation mappings are model-derived. Use as a research exploration tool, not for clinical decision-making.

Examples

## Not run: 
summary <- mskKGSummary()
print(summary)

## End(Not run)

Description

Computes per-timepoint RMS activation, synergy decomposition (W/H/VAF), and optionally CMC for a series of EMG measurements taken at different timepoints during rehabilitation.

Usage

mskLongitudinalTracker(
  timepoints,
  hg = NULL,
  emg_mapping = NULL,
  metrics = c("rms", "synergy"),
  sr = NULL
)

Arguments

Value

An S3 object of class "MSKLongitudinalTracker" with:

metrics_table

data.frame (timepoint, muscle, metric_name, value)

timepoint_labels

character vector

n_timepoints

integer

synergy_series

list of synergy results per timepoint

activation_series

matrix (n_muscles x n_timepoints)

Examples

## Not run: 
emg_t0 <- matrix(abs(rnorm(400)), 100, 4)
emg_t1 <- matrix(abs(rnorm(400)), 100, 4)
colnames(emg_t0) <- colnames(emg_t1) <- paste0("muscle_", 1:4)
tracker <- mskLongitudinalTracker(list(T0 = emg_t0, T1 = emg_t1))

## End(Not run)

Description

Computes ICC (intraclass correlation coefficient) across timepoints for each metric, then derives SEM and MDC values for clinimetric assessment.

Usage

mskMinimalDetectableChange(
  tracker,
  method = "standard",
  confidence = 0.95,
  icc_method = "ICC(2,1)"
)

Arguments

Value

An S3 object of class "MSKMinimalDetectableChange" with:

mdc_table

data.frame (muscle, metric, ICC, SEM, MDC, pooled_SD)

confidence_level

numeric

method

character

Examples

## Not run: 
mdc <- mskMinimalDetectableChange(tracker, confidence = 0.95)

## End(Not run)

Description

Computes the modularity Q of a partition with resolution parameter. Q = (1/2m) * sum_ij (A_ij - gamma * k_i * k_j / (2m)) * delta(c_i, c_j)

Usage

mskModularity(A, membership, gamma = 1)

Arguments

Value

Numeric modularity value.

Description

Computes standard graph metrics for the projected bone or muscle graph.

Usage

mskNetworkMetrics(hg, type = c("bone", "muscle"))

Arguments

Value

A list with: degree, strength, clustering_coef, density

Description

Compares corticomuscular coherence (CMC) and directional coupling between two timepoints to quantify neural adaptation during rehabilitation.

Usage

mskNeuralAdaptationIndex(
  cmc_t0,
  cmc_t1,
  directional_t0 = NULL,
  directional_t1 = NULL
)

Arguments

Value

A list with: cmc_change, directional_change, adaptation_index, interpretation ("improving"/"stable"/"declining").

Examples

## Not run: 
nai <- mskNeuralAdaptationIndex(cmc_t0, cmc_t1, dir_t0, dir_t1)

## End(Not run)

Description

Creates multiple null hypergraphs and computes impact scores for each. This is used to compute impact deviations relative to the null distribution.

Usage

mskNullEnsemble(hg = NULL, n_null = 100L, sim_params = list(), verbose = TRUE)

Arguments

Value

A list with:

null_scores

Matrix of impact scores (n_muscles x n_null)

null_degree_scores

List, impact scores grouped by degree for each null

n_null

Number of null models

References

Murphy AC et al. (2018) PLOS Biology.

Description

Creates a randomized version of the musculoskeletal hypergraph by rewiring muscle-bone connections while preserving each muscle's degree (number of bones it connects to). This is done by randomly reassigning which bones each muscle attaches to, within anatomical categories if specified.

Usage

mskNullHypergraph(hg = NULL, preserve_category = FALSE)

Arguments

Value

An MSKHypergraph object with rewired connections.

References

Murphy AC et al. (2018) PLOS Biology.

Description

Computes confidence intervals for functional outcome predictions using bootstrap resampling or analytical delta method.

Usage

mskOutcomeConfidenceInterval(
  predictions,
  method = c("bootstrap", "analytical"),
  n_boot = 100L,
  confidence_level = 0.95
)

Arguments

Value

A list with:

ci

data.frame with outcome, point_estimate, lower, upper, width

method

character

n_boot

integer (for bootstrap)

overall_uncertainty

numeric, mean CI width

Examples

## Not run: 
outcome <- mskPredictFunctionalOutcome("Biceps Brachii")
ci <- mskOutcomeConfidenceInterval(outcome)

## End(Not run)

Description

Orchestrator function that combines functional outcome prediction, milestones, and optional reassessment into a comprehensive report.

Usage

mskOutcomeReport(outcome, milestones = NULL, reassessment = NULL, hg = NULL)

Arguments

Value

An S3 object of class "MSKOutcomeReport" with all sub-results.

Examples

## Not run: 
outcome <- mskPredictFunctionalOutcome("Biceps Brachii")
milestones <- mskFunctionalMilestones(outcome)
report <- mskOutcomeReport(outcome, milestones)
print(report)

## End(Not run)

Description

Orchestrator function that calls all clinical bridge functions with shared hypergraph and simulation objects, producing a comprehensive report.

Usage

mskOutcomeSummary(injury_muscles, patient_data = NULL, hg = NULL, sim = NULL)

Arguments

Value

An S3 object of class "MSKOutcomeSummary" containing:

prediction

MSKClinicalPrediction object

timeline

Recovery timeline data.frame

risk_profile

MSKInjuryRiskProfile object (if patient_data provided)

rehab

MSKRehabProtocol object

Clinical Validity

All predictions are model-based estimates using MSK network topology. Recovery model was validated on 14 aggregate muscle groups. Patient factors are heuristic. This is a research exploration tool, not a clinical diagnostic.

Examples

## Not run: 
summary <- mskOutcomeSummary("Trapezius")
print(summary)

## End(Not run)

Description

Finds the shortest anatomical pathway between two entities (muscles, bones, nerves, etc.) in the knowledge graph. Useful for tracing innervation chains, biomechanical linkages, and anatomical relationships.

Usage

mskPathwayQuery(from, to, hub = NULL, max_depth = 5L)

Arguments

Value

A list with:

path

Character vector of entities along the path

predicates

Character vector of relationship types between entities

depth

Integer, path length

description

Human-readable path description

found

Logical, whether a path was found

Clinical Validity

Pathways reflect relationships encoded in the knowledge graph. The shortest path in the KG may not correspond to the most clinically relevant connection. Always verify pathway interpretations against anatomical references.

Examples

## Not run: 
path <- mskPathwayQuery("Biceps Brachii", "C5")
cat(path$description, "\n")

## End(Not run)

Description

Predicts clinical functional outcomes (ROM, strength, composite function score) for injured muscles based on MSK network topology, impact analysis, and optional EMG activation data. Provides evidence-based regression models with confidence intervals.

Usage

mskPredictFunctionalOutcome(
  injured_muscles,
  hg = NULL,
  outcome_type = c("all", "rom", "strength", "function"),
  patient_factors = NULL,
  emg = NULL,
  emg_mapping = NULL,
  confidence_level = 0.95
)

Arguments

Value

An S3 object of class "MSKFunctionalOutcome" with:

predictions

data.frame with muscle, outcome_type, predicted_value, lower_ci, upper_ci, unit

aggregate

list with overall_rom, overall_strength, overall_function

recovery_weeks

estimated weeks to reach 90 percent of predicted outcome

confidence_level

numeric

model_type

character

patient_factors_used

list

Clinical Validity

The prediction models are based on MSK network topology and heuristic adjustments. They are not independently validated clinical tools. Use as a research exploration tool, not a clinical diagnostic.

Examples

## Not run: 
outcome <- mskPredictFunctionalOutcome(
  c("Biceps Brachii", "Deltoid"),
  patient_factors = list(age = 55, injury_severity = "moderate")
)
print(outcome)

## End(Not run)

Description

Compares current physiological measurements to a previous assessment, computing progress metrics and generating recommendations.

Usage

mskReassess(
  current_data,
  previous_assessment,
  hg = NULL,
  emg_mapping = NULL,
  sr = NULL
)

Arguments

Value

An S3 object of class "MSKReassessment" with:

current_metrics

data.frame of current activation/synergy metrics

previous_metrics

data.frame from previous assessment

change

data.frame with metric, previous, current, change, pct_change, significant

progress_summary

character vector per metric

overall_progress

weighted average progress score (0-100)

recommendation

character

Examples

## Not run: 
reassessment <- mskReassess(
  current_data = list(emg = emg_matrix),
  previous_assessment = outcome
)
print(reassessment)

## End(Not run)

Description

Generates a week-by-week recovery timeline based on exponential decay of network impact, with clinical phase assignments.

Usage

mskRecoveryTimeline(injury_muscles, hg = NULL, sim = NULL, n_weeks = 12L)

Arguments

Value

A data.frame with columns: week, muscle, remaining_impact_pct, phase, milestone.

Clinical Validity

Phase assignments are based on exponential decay of network impact scores, not empirical clinical data. They should be interpreted as model-based estimates for research purposes only.

Examples

## Not run: 
timeline <- mskRecoveryTimeline("Biceps Brachii")

## End(Not run)

Description

Fits parametric recovery curves (exponential, sigmoid, or linear) to longitudinal muscle activation data from an MSKLongitudinalTracker.

Usage

mskRecoveryTrajectoryFit(
  tracker,
  model = c("exponential", "sigmoid", "linear"),
  muscle_subset = NULL
)

Arguments

Value

An S3 object of class "MSKRecoveryTrajectory" with:

fits

list per muscle with coefficients, residuals, R-squared, AIC

predicted

matrix of predicted values (muscles x timepoints)

model_type

character

recovery_rate

numeric vector (slope at midpoint for each muscle)

time_to_90pct

estimated time to 90 percent of asymptotic recovery

Examples

## Not run: 
traj <- mskRecoveryTrajectoryFit(tracker, model = "exponential")

## End(Not run)

Description

Generates a phased rehabilitation protocol using MSK network community structure and shortest path distances to determine exercise progression.

Usage

mskRehabProtocol(injury_muscles, hg = NULL, n_phases = 3L)

Arguments

Value

An S3 object of class "MSKRehabProtocol" with per-phase muscle lists.

Clinical Validity

Phase ordering is based on network topology (community membership and shortest path distance), not validated rehabilitation protocols. Use as a research exploration tool, not clinical guidance.

Examples

## Not run: 
protocol <- mskRehabProtocol("Biceps Brachii")
print(protocol)

## End(Not run)

Description

Master function that runs the complete analysis pipeline and compares results to the paper's reported values. This is the main validation function.

Usage

mskReproducePaper(run_simulation = FALSE, verbose = TRUE)

Arguments

Value

A list with all results and paper comparisons.

Description

Fits a robust linear model (using MASS::rlm if available, otherwise lm). Supports weighted regression as used in the paper.

Usage

mskRobustRegression(x, y, weights = NULL)

Arguments

Value

A list with:

coefficients

Named vector (intercept, slope)

r_squared

R-squared value

f_statistic

F-statistic

p_value

p-value for the regression

residuals

Residual values

fitted

Fitted values

model

The fitted model object

Description

Computes shortest path distance matrix for the projected graph.

Usage

mskShortestPaths(hg, type = c("bone", "muscle"))

Arguments

Value

A numeric matrix of shortest path distances.

Description

Runs a damped harmonic oscillator simulation on the musculoskeletal network. Each muscle is modeled as a spring connecting its attached bones, and each bone is a unit-mass point particle. The system is evolved under perturbation and the total displacement is computed as the impact score.

Usage

mskSimulate(
  hg = NULL,
  dt = 0.01,
  n_steps = 500L,
  beta = 1,
  perturbation_magnitude = 1
)

Arguments

Value

An S3 object of class "MSKSimulation" with:

hg

The MSKHypergraph used

spring_constants

Named vector of spring constants per muscle

params

List of simulation parameters

References

Murphy AC et al. (2018) PLOS Biology 16(1): e2002811.

Description

Compares two synergy decompositions by aligning synergy weight vectors and computing a change index reflecting structural reorganization.

Usage

mskSynergyChangeIndex(
  synergy_t0,
  synergy_t1,
  method = c("cosine", "correlation", "procrustes")
)

Arguments

Value

A list with: global_change_index (0=identical, 1=maximally different), per_synergy_change, alignment (permutation used).

Examples

## Not run: 
sci <- mskSynergyChangeIndex(synergy_t0, synergy_t1, method = "cosine")

## End(Not run)

Description

Computes the z-scored Rand index between two partitions. Used to compare community structure with homunculus categories.

Usage

mskZRand(partition1, partition2)

Arguments

Value

Numeric z-Rand score. Values > 1.96 indicate significant similarity.

References

Traud et al. (2011) Physical Review E.

Description

Computes pairwise corticomuscular coherence (CMC) between EEG and EMG channels, maps EMG channels to MSK muscles, builds a CMC-based functional distance matrix, and compares it with the structural muscle adjacency via Mantel test.

Usage

neuromechCorticomuscularCoupling(
  eeg,
  emg,
  hg = NULL,
  freq_band = c(15, 35),
  eeg_channels = NULL,
  emg_mapping = NULL,
  sr_eeg = NULL,
  sr_emg = NULL,
  nperseg = 256L,
  n_perm = 999L
)

Arguments

Value

An S3 object of class "MSKNeuromechCMC" with:

cmc_matrix

Coherence matrix (n_eeg x n_emg)

structural_matrix

MSK muscle adjacency (matched subset)

emg_cmc_profile

Per-muscle mean CMC across EEG channels

significant_pairs

Data.frame of significant EEG-EMG pairs

mantel

Mantel test result (correlation, p_value)

mapping

EMG-to-muscle mapping used

Examples

## Not run: 
result <- neuromechCorticomuscularCoupling(eeg_data, emg_data)

## End(Not run)

Description

Computes directional (causal) connectivity between EEG and EMG channels using Granger causality or transfer entropy, distinguishing descending (cortical->muscle) from ascending (proprioceptive) pathways.

Usage

neuromechDirectionalCoupling(
  eeg,
  emg,
  hg = NULL,
  method = c("granger", "transfer_entropy", "both"),
  max_order_ms = 50,
  lag_ms = 20,
  n_bins = NULL,
  eeg_channels = NULL,
  emg_mapping = NULL,
  sr_eeg = NULL,
  sr_emg = NULL,
  n_perm = 999L,
  alpha = 0.05
)

Arguments

Value

An S3 object of class "MSKNeuromechDirectional" with:

descending

Matrix (n_eeg x n_emg): EEG->EMG values

ascending

Matrix (n_emg x n_eeg): EMG->EEG values

net_direction

descending - t(ascending)

significant_descending

Data.frame of significant descending pairs

significant_ascending

Data.frame of significant ascending pairs

dominance_ratio

Per-muscle mean(descending)/mean(ascending)

pathway_classification

Data.frame classifying each pair

msk_correlation

Mantel test result

method

Method used

parameters

List of parameters used

Examples

## Not run: 
result <- neuromechDirectionalCoupling(eeg_data, emg_data, method = "granger")

## End(Not run)

Description

Computes the electromechanical delay (EMD) between EMG and kinematic signals for muscle-bone pairs connected in the MSK hypergraph. Optionally correlates EMD with MSK network distance.

Usage

neuromechElectromechanicalDelay(
  emg,
  kinematics,
  hg = NULL,
  emg_mapping = NULL,
  kin_mapping = NULL,
  sr = NULL,
  max_lag_ms = 200,
  window_sec = NULL,
  n_perm = 999L
)

Arguments

Value

A list with:

emd

Data.frame with muscle, bone, emd_ms, peak_correlation per pair

network_distance

Corresponding MSK shortest path distance

correlation

Pearson correlation between EMD and network distance

p_value

Permutation p-value for the correlation

time_varying

Data.frame with per-window EMD (if window_sec set)

Examples

## Not run: 
result <- neuromechElectromechanicalDelay(emg_data, kin_data)

## End(Not run)

Description

Combines EMG activation, kinematic stress, and force data to compute a multi-source vulnerability score for each muscle. Propagates kinematic stress through the MSK incidence matrix and weights by impact deviation.

Usage

neuromechIntegratedVulnerability(
  emg,
  kinematics,
  force_data,
  hg = NULL,
  emg_mapping = NULL,
  kin_mapping = NULL,
  weights = c(1, 1, 1),
  use_proxy = TRUE
)

Arguments

Value

A list with:

vulnerability

Named numeric vector of muscle vulnerability scores

ranking

Data.frame sorted by vulnerability (descending)

source_correlations

Pairwise correlations between EMG, kin, force sources

Examples

## Not run: 
result <- neuromechIntegratedVulnerability(emg, kinematics, force)

## End(Not run)

Description

Estimates joint torques from EMG activation and anatomical moment arms, computing per-muscle contributions, coactivation indices, and torque balance ratios across joints.

Usage

neuromechJointTorque(
  emg,
  hg = NULL,
  emg_mapping = NULL,
  moment_arm_table = NULL,
  activation_method = c("rms", "mean_rectified", "peak"),
  sr = NULL,
  joints = NULL,
  n_perm = 999L
)

Arguments

Value

An S3 object of class "MSKNeuromechTorque" with:

per_muscle

Data.frame: muscle, joint, activation, moment_arm, direction, torque_contribution

per_joint

Data.frame: joint, net_torque, agonist_sum, antagonist_sum, coactivation_index, n_muscles

torque_balance

Data.frame: joint, balance_ratio

msk_correlation

Mantel test result

moment_arm_source

Character: "lookup" or "custom"

Examples

## Not run: 
result <- neuromechJointTorque(emg_data, hg = hg)

## End(Not run)

Description

Maps cortical drive (CMC), muscle activation (EMG RMS), and force output to the MSK hypergraph to compute per-muscle drive efficiency and aggregate per community.

Usage

neuromechMotorDriveTopography(
  eeg,
  emg,
  force_data,
  hg = NULL,
  freq_band = c(15, 35),
  gamma = 4.3,
  emg_mapping = NULL,
  sr = NULL
)

Arguments

Value

A list with:

per_muscle

Data.frame with muscle, cortical_drive, emg_activation, force, drive_efficiency

per_community

Data.frame with community, mean drive/activation/force

drive_force_correlation

Pearson r between cortical drive and force

drive_force_p_value

p-value for the correlation

Examples

## Not run: 
result <- neuromechMotorDriveTopography(eeg, emg, force)

## End(Not run)

Description

Extracts muscle synergies from EMG data via NMF or PCA, maps them to MSK structural communities, and compares synergy-based partitioning with structural communities via z-Rand.

Usage

neuromechMuscleSynergy(
  emg,
  hg = NULL,
  emg_mapping = NULL,
  method = c("nmf", "pca"),
  n_synergies = 4L,
  auto_select = FALSE,
  vaf_threshold = 0.9,
  max_k = 10L,
  gamma = 4.3,
  n_perm = 999L,
  seed = NULL,
  sr = NULL
)

Arguments

Value

An S3 object of class "MSKNeuromechSynergy" with:

W

Synergy weight matrix (n_muscles x n_synergies)

H

Activation coefficients (n_synergies x n_timepoints)

n_synergies

Number of synergies extracted

vaf

Variance accounted for

vaf_curve

VAF curve if auto_select (else NULL)

method

Decomposition method used

community_mapping

Synergy-to-community enrichment data.frame

synergy_similarity

Cosine similarity matrix (k x k)

community_synergy_zrand

z-Rand comparing synergy vs structural

reconstruction_error

Reconstruction error

Examples

## Not run: 
result <- neuromechMuscleSynergy(emg_data, method = "nmf", n_synergies = 4)

## End(Not run)

Description

Orchestrator function that runs all available neuromechanics analyses with shared mapping, skipping analyses when inputs are NULL.

Usage

neuromechSummary(
  eeg = NULL,
  emg,
  kinematics = NULL,
  force_data = NULL,
  hg = NULL,
  freq_band = c(15, 35),
  gamma = 4.3,
  sr = NULL,
  n_synergies = 4L,
  synergy_method = "nmf"
)

Arguments

Value

An S3 object of class "MSKNeuromechSummary" with:

cmc

CMC result or NULL

emd

EMD result or NULL

motor_drive

Motor drive result or NULL

vulnerability

Vulnerability result or NULL

torque

Joint torque result or NULL

synergy

Muscle synergy result or NULL

directional

Directional coupling result or NULL

available_analyses

Character vector of successfully computed analyses

Examples

## Not run: 
summary <- neuromechSummary(eeg = eeg_data, emg = emg_data,
                             kinematics = kin_data, force_data = force)

## End(Not run)

Description

Runs network analysis on multiple OpenSim models and performs statistical comparisons of their network properties.

Usage

opensimCompareModels(model_paths, names = NULL, gamma = 4.3)

Arguments

Value

An S3 object of class "MSKModelComparison" with:

analyses

List of MSKOpenSimAnalysis objects

comparisons

Data frame of pairwise comparison statistics

Description

Combines muscle force data from OpenSim simulations with MSK network impact scores to compute force-weighted vulnerability.

Usage

opensimForceToImpact(
  model_path,
  force_data,
  hg = NULL,
  method = c("weighted_simulation", "weighted_scores")
)

Arguments

Value

A list with:

weighted_impact_scores

Force-weighted impact scores

force_weights

Force values used for weighting

unweighted_scores

Original (unweighted) impact scores

Description

Builds a hypergraph from an OpenSim model and runs complete network analysis including metrics, community detection, and optionally impact scoring.

Usage

opensimNetworkAnalysis(
  model = NULL,
  model_path = NULL,
  gamma = 4.3,
  run_simulation = TRUE
)

Arguments

Value

An S3 object of class "MSKOpenSimAnalysis" with:

hypergraph

The MSKHypergraph object

metrics

Network metrics from mskNetworkMetrics

communities

Community detection results

impact_scores

Impact scores (if run_simulation = TRUE)

Description

Parses an OpenSim .osim file to extract muscle-bone attachment points and constructs a subject-specific MSKHypergraph.

Usage

opensimToMSKHypergraph(model = NULL, model_path = NULL)

Arguments

Value

An MSKHypergraph object with subject-specific anatomy.

Note

Requires the xml2 package.

Examples

## Not run: 
hg <- opensimToMSKHypergraph(model_path = "gait2392.osim")
print(hg)

## End(Not run)

Description

Visualizes detected communities with size information.

Usage

plotCommunityStructure(community, ...)

Arguments

Value

Invisible NULL.

Description

Plots the degree distribution of the hypergraph, comparing real and null model distributions. Reproduces Fig 2e.

Usage

plotDegreeDistribution(
  hg = NULL,
  type = c("muscle", "bone"),
  null_dist = NULL,
  log_scale = TRUE,
  ...
)

Arguments

Value

Invisible NULL.

Description

Recreates Fig 4b: deviation ratio vs homunculus area. Target: F(1,19) = 21.3, R^2 = 0.52

Usage

plotHomunculus(homunculus_data = NULL, ...)

Arguments

Value

Invisible the regression result.

Description

Recreates Fig 3a: scatter plot of impact scores by hyperedge degree. Target: R^2 = 0.45 for the relationship.

Usage

plotImpactVsDegree(impact_scores, hg = NULL, show_regression = TRUE, ...)

Arguments

Value

Invisible the regression result.

Description

Recreates Fig 3b: weighted regression of impact deviation vs clinical recovery time. Target: R^2 = 0.757.

Usage

plotImpactVsRecovery(recovery_data = NULL, ...)

Arguments

Value

Invisible the regression result.

Description

Visualizes the musculoskeletal network with optional community coloring.

Usage

plotMSKNetwork(
  hg = NULL,
  type = c("muscle", "bone"),
  membership = NULL,
  layout = "fr",
  vertex_size = 3,
  ...
)

Arguments

Value

Invisible NULL. Produces a plot.

Description

Creates the bone-bone weighted adjacency matrix A = t(C) %*% C where A[i,j] counts the number of muscles shared between bones i and j. This is the one-mode projection onto the bone (vertex) space.

Usage

projectBoneGraph(hg)

Arguments

Value

A symmetric sparse Matrix (n_bones x n_bones)

References

Murphy AC et al. (2018) PLOS Biology.

Description

Creates the muscle-muscle weighted adjacency matrix B = C %*% t(C) where B[i,j] counts the number of bones shared between muscles i and j. This is the one-mode projection onto the muscle (hyperedge) space.

Usage

projectMuscleGraph(hg)

Arguments

Value

A symmetric sparse Matrix (n_muscles x n_muscles)

References

Murphy AC et al. (2018) PLOS Biology.

Description

Returns the degree of each vertex (bone), i.e., the number of muscles attached to each bone.

Usage

vertexDegree(hg)

Arguments

Value

Named numeric vector of bone degrees.