Introduction: Prior studies comparing Hemodynamic Response Function (HRF) in the young
and elderly adults based on fMRI data have reported inconsistent findings for brain vision and
motor regions in healthy aging. It is shown that the averaging method employed in all previous
works has caused this inconsistency. The averaging is so sensitive to outliers and noise. However,
fMRI data are obscured with a major contribution of noise particularly in the elderly case.
Methods: Deconvolution algorithm is here proposed for HRF extraction to achieve more
robustness against noise. In spite of earlier works, proposed deconvolution algorithm yields
compatible HRF results using either original or denoised fMRI data, though a large percentage
of selected active voxels change in the latter case. In the current study, event-related fMRI
data have been used for 18 subjects (8 young and 10 elderly adults) with a simple visual and
motor task of pressing a key with index in response to the visual presentation of the word tap.
Considering anatomically-defined vision and motor regions and preprocessing steps in FSL and
SPM, the activated voxels have been selected according to t-test for which HRF is estimated
using deconvolution method.
Results: Experimental results demonstrate that HRF peak amplitudes do not differ significantly
(p=0.8) in the vision region for the young and the elderly. In motor region, the HRF peak
significantly increases for the young compared to the elderly (p<0.03). Repeating the procedure
on the denoised fMRI data using MDL algorithm, the same results have been obtained.
Discussion: In this study, a comparative study has been realized on the hemodynamic
response properties associated with the young and the elderly adults on a simple visual
and motor task.
Type of Study:
Original |
Subject:
Computational Neuroscience Received: 2014/04/5 | Accepted: 2014/11/11 | Published: 2015/01/1