Table 1.
Demographic, retinal, and MRI measures in study participants.
When p values were not listed they were at least > 0.4. WM–white matter, CRAE–central retinal arterial equivalents, CRVE–central retinal venous equivalents. FA–fractional anisotropy, MD–mean diffusivity, All fractal dimension values are dimensionless. A straight line has a fractal dimension of 1, and a plane has a fractal dimension of 2. FA and MD values range between 0 and 1 and are dimensionless. CRAE and CRVE are in microns. CRAE/CRVE is dimensionless. All volumes (total gray, subcortical volumes, and log of WM hyperintensities) are in milliliters. For continuous variables (i.e. all variables except number with hypertension, and number of African Americans) values listed are means with the standard deviation in parentheses. CD4 number is cells per microliter.
Table 2.
Regression models of the association of fractal dimension with fractional anisotropy.
The predictors included in the model are listed in the first column, the second column lists the predictor of interest, and the third column lists the standardized β coefficients and p values. ICV = intracranial volume. Fractal anisotropy (FA) was the dependent variable for all analyses in the table. The standardized beta coefficient for arterial dimension is this model was 0.17, p = 0.21.
Table 3.
Associations of arterial and venous retinal fractal dimension with brain outcomes.
N = 55. Values listed are standardized β-coefficients and significance levels from linear regression models. All models included both arterial and venous dimension in addition to the variables listed in the top row. AA–African-American race/ethnicity, HIV–HIV-infected, ICV–intracranial volume. FA–fractional anisotropy, MD–mean diffusivity, WM–white matter.
Table 4.
Association of age and blood pressure with fractal dimension.
Values listed are standardized beta coefficients and significance levels for linear regression models predicting either mean arterial dimension or mean venous dimension where age, systolic, and diastolic blood pressures were independent variables.