Skip to main content
Fig. 1 | The Journal of Headache and Pain

Fig. 1

From: Machine learning approach for Migraine Aura Complexity Score prediction based on magnetic resonance imaging data

Fig. 1

Correlation analysis of features that significantly correlate with MACS (< 0.05). The correlation-based feature selection includes the elimination of features that have a correlation greater than ±0.85. Correlation values range from -1 to 1 as shown in the heatmap color spectrum. The purple color symbolises a negative correlation, the middle shades indicate a low correlation, whereas the orange section indicates a positive correlation. The color intensity shows the correlation strength. (1 - Left parahippocampal gyrus mean Gaussian curvature; 2 - Left transverse temporal gyrus mean Gaussian curvature; 3 - Left transverse temporal gyrus thickness; 4 - Left pars opercularis thickness; 5 - Left lingual gyrus surface area; 6 - Right transverse temporal gyrus mean Gaussian curvature; 7 - Left caudal middle frontal gyrus surface area; 8 - Left entorhinal gyrus volume; 9 - Right entorhinal gyrus volume; 10 - Right lingual gyrus surface area; 11 - Left precuneus gyrus surface area; 12 - Left pars opercularis surface area; 13 - Right caudal middle frontal gyrus surface area; 14 - Right transverse temporal gyrus thickness; 15 - Left caudal middle frontal gyrus volume; 16 - Left isthmus cingulate surface area; 17 - Left pericalcarine thickness; 18 - Left caudal anterior cingulate mean Gaussian curvature; 19 - Right entorhinal gyrus surface area; 20 - Right pericalcarine mean Gaussian curvature; 21 - Left medial orbitofrontal gyrus folding index; 22 - Left pars triangularis thickness; 23 - Right caudal middle frontal gyrus folding index; 24 - Right paracentral gyrus mean Gaussian curvature; 25 - Left lingual gyrus volume; 26 - Right lingual gyrus volume)

Back to article page