Post-stroke deficit prediction from lesion and indirect structural and functional disconnection

Abstract

Behavioural deficits in stroke reflect both structural damage at the site of injury, and widespread network dysfunction caused by structural, functional, and metabolic disconnection. Two recent methods allow for the estimation of structural and functional disconnection from clinical structural imaging. This is achieved by embedding a patient's lesion into an atlas of functional and structural connections in healthy subjects, and deriving the ensemble of structural and functional connections that pass through the lesion, thus indirectly estimating its impact on the whole brain connectome. This indirect assessment of network dysfunction is more readily available than direct measures of functional and structural connectivity obtained with functional and diffusion MRI, respectively, and it is in theory applicable to a wide variety of disorders. To validate the clinical relevance of these methods, we quantified the prediction of behavioural deficits in a prospective cohort of 132 first-time stroke patients studied at 2 weeks post-injury (mean age 52.8 years, range 22-77; 63 females; 64 right hemispheres). Specifically, we used multivariate ridge regression to relate deficits in multiple functional domains (left and right visual, left and right motor, language, spatial attention, spatial and verbal memory) with the pattern of lesion and indirect structural or functional disconnection. In a subgroup of patients, we also measured direct alterations of functional connectivity with resting-state functional MRI. Both lesion and indirect structural disconnection maps were predictive of behavioural impairment in all domains (0.16 < R2 < 0.58) except for verbal memory (0.05 < R2 < 0.06). Prediction from indirect functional disconnection was scarce or negligible (0.01 < R2 < 0.18) except for the right visual field deficits (R2 = 0.38), even though multivariate maps were anatomically plausible in all domains. Prediction from direct measures of functional MRI functional connectivity in a subset of patients was clearly superior to indirect functional disconnection. In conclusion, the indirect estimation of structural connectivity damage successfully predicted behavioural deficits post-stroke to a level comparable to lesion information. However, indirect estimation of functional disconnection did not predict behavioural deficits, nor was a substitute for direct functional connectivity measurements, especially for cognitive disorders.

Keywords: connectivity; functional disconnection; lesion; stroke; structural disconnection.

Figure 1

Schematic flowchart of analysis. Individual lesion, SDC, and FDC maps are used as dependent variables for predicting behavioural scores using ridge regression with leave-one-out cross validation. The output of the ridge regression is a map of the projected weights for each input, and an estimate of the best possible predicted variance between real and predicted scores (R² and scatter plot).

Figure 2

Ridge regression model for right visual pattern score (left hemisphere lesion, 31 subjects) for lesion, SDC and FDC. Predictive weights projected back on the brain (left), and plot of the real score (x-axis) versus model predicted score (y-axis) (right) are reported. R²-value is reported on the scatter plot. Red-yellow represent voxels predicting deficits while blue-green represent voxels predicting no deficit. To optimize the visualization, the normalized projected values in the range (−0.1,0.1) are not displayed.

Figure 3

Ridge regression model for left motor scores (right hemisphere lesion, 51 subjects) for lesion, SDC and FDC. Predictive weights projected back on the brain (left), and plot of the real score (x-axis) versus model predicted score (y-axis) (right) are reported. R²-value is reported on the scatter plot. Red-yellow represent voxels predicting deficits while blue-green represent voxels predicting no deficit. To optimize the visualization, the normalized projected values in the range (−0.1,0.1) are not displayed.

Figure 4

Ridge regression model for language factorial score (116 subjects) for lesion, SDC and FDC. Predictive weights projected back on the brain (left), and plot of the real score (x-axis) versus model predicted score (y-axis) (right) are reported. R-value is reported on the scatter plot. Red-yellow represent voxels predicting deficits while blue-green represent voxels predicting no deficit. To optimize the visualization, the normalized projected values in the range (−0.1,0.1) are not displayed.

Figure 5

Ridge regression model for verbal memory factorial score (88 subjects) for lesion, SDC and FDC. Predictive weights projected back on the brain (left), and plot of the real score (x-axis) versus model predicted score (y-axis) (right) are reported. R²-value is reported on the scatter plot. Red-yellow represent voxels predicting deficits while blue-green represent voxels predicting no deficit. To optimize the visualization, the normalized projected values in the range (−0.1,0.1) are not displayed.

Figure 6

Ridge regression model on language score (88 subjects) for SDC, FDC and r-fMRI FC changes. Predictive weights projected back on the brain (left) and plot of the real score (x-axis) and model predicted score (y-axis) (right) are reported. R²-value is reported on the scatter plot. In upper and central maps, red-yellow represent voxels predicting deficits while blue-green represent voxels predicting no deficit. To optimize the visualization, the projected values in range (−0.1,0.1) are not displayed. Maps of r-fMRI FC (bottom) changes represent most predictive connections and nodes for FC-deficit model. The top 200 connections are shown: green connections indicate positive projected values (better performance) and orange connections indicates negative projected values (worse performance). The subset of 324 parcels included in the top 200 connections are displayed, the size of nodes is related to their contribution to the model (calculated as root-mean-square of all connections for each node).

Figure 7

Ridge regression model for spatial memory score (71 subjects) for SDC, FDC and r-fMRI FC changes. Predictive weights projected back on the brain (left) and plot of the real score (x-axis) and model predicted score (y-axis) (right) are reported. R²-value is reported on the scatter plot. In upper and central maps, red-yellow represent voxels predicting deficits while blue-green represent voxels predicting no deficit. To optimize the visualization, the projected values in range (−0.1,0.1) are not displayed. Maps of r-fMRI FC (bottom) changes represent most predictive connections and nodes for FC-deficit model. The top 200 connections are shown: green connections indicate positive projected values (better performance) and orange connections indicates negative projected values (worse performance). The subset of 324 parcels included in the top 200 connections are displayed, the size of nodes is related to their contribution to the model (calculated as root-mean-square of all connections for each node).