Keywords:
Neuroradiology brain, MR, Experimental investigations, Dementia, Geriatrics
Authors:
P. Naumczyk1, A. K. sawicka1, B. Brzeska1, A. Sabisz1, K. Jodzio1, P. J. Winklewski1, K. Finc2, E. Szurowska1, A. Szarmach1; 1Gdansk/PL, 2Toruń/PL
DOI:
10.26044/ecr2019/C-0797
Methods and materials
Participants
Forty-six older healthy subjects (37 women,
9 men) over 60 years of age were included in the study.
The participants were examined on two 3.0T MRI scanners (Philips Achieva TX and GE Discovery 750).
Methods
The protocol included standard anatomical sequences (to exclude brain pathology) as well as a high-resolution T1-weighted sequence used later for cortical thickness estimation (Philips: repetition time [TR]=8,1ms,
echo time [TE]=3,7ms,
voxel size 1x1x1mm,
flip angle: 8o; GE: TR=8,192ms,
TE=3,18ms,
voxel size 1x1x1mm,
flip angle=12o).
The neuropsychological protocol included fluid intelligence assessment (Raven Progressive Matrices standard version),
crystalized intelligence assessment (Information subtest of the Wechsler Adult Intelligence Scale – Revised [WAIS-R] or Vocabulary subtest of the WAIS-R),
and executive functioning (Colour Traits Test).
Demographical and neuropsychological characteristics of the group is presented in Table 1.
Cortical thickness measurements were conducted in FreeSurfer software with Desikan-Killany brain atlas as a reference (Figure 1).
Statistical analyses
Statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) software,
version 24 (IBM,
New York,
US).
The analyses focused on the investigation of the relationship between the obtained neuropsychological metrices (fluid IQ raw score,
crystallized IQ calculated normative score,
time for part 1 of the CTT,
time for part 2 of the CTT,
and the interference index of CTT),
age,
and the cortical thickness scores derived from the MRI examination.
The normality of the variables’ distribution was examined and then the analyses included paired correlations between all the variables with Pearson’s and Spearman’s tests (depending on whether the given variable was distributed normally).
Afterwards for all anatomical ROIs,
for which more than one of the neuropsychological/demographical variables correlated with cortical thickness,
a hierarchical stepwise regression (for normally distributed variables),
or a partial Spearman’s correlation (for non-normally distributed variables) was performed to determine which of the examined factors explained most of the variance of grey matter thickness.