Age-related degradation of optic radiation
white matter predicts visual, but not verbal
executive functions
Christina E. Webb, Patricio M. Viera Perez, David A. Hoagey, Chen Gonen, Karen M. Rodrigue, Kristen M. Kennedy
Center for Vital Longevity, School of Behavioral and Brain Sciences, The University of Texas at Dallas
Webb etal. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 IGO License, which permits the copy
and redistribution of the material in any medium or format provided the original work and author are properly credited. In any reproduction of this article there should not be any
suggestion that APERTURE NEURO or this article endorse any specic organization or products. The use of the APERTURE NEURO logo is not permitted. This notice should be
preserved along with the article’s original URL. Open access logo and text by PLoS, under the Creative Commons Attribution-Share Alike 4.0 Unported license.
: 2022, Volume 2 - 1 - CC-BY: © Webb etal.
Healthy aging is accompanied by degraded white matter connectivity, which has been suggested to contribute to cognitive dys-
function observed in aging, especially in relation to uid measures of cognition. Prior research linking white matter microstructure
and cognition, however, has largely been limited to major association and heteromodal white matter tracts. The optic radiations
(OR), which transfer visual sensory-perceptual information from thalamic lateral geniculate nucleus to primary visual cortex, are gen-
erally considered lower-level input-relay white matter tracts. However, the role of this prominent white-matter visual relay system
in supporting higher-order cognition is understudied, especially in regard to healthy aging. The present study used deterministic
tractography to isolate OR fractional anisotropy (FA) in 130 highly screened, healthy participants aged 20–94 years (mean 47.88 ±
17.36 SD; 76 women) to assess age effects on OR tract white matter. We also examined associations between age-related differences
in the OR and cognitive domains involving visual processing speed, and visual- and non-visual executive function (EF). OR microstruc-
ture, as indexed by FA, exhibited a signicant linear decrease across age. A signicant interaction between age, FA, and cognitive do-
main on cognitive task performance indicated that in older age, more degraded OR white matter was associated with poorer visual EF,
but no age-related association between FA in the OR and visual processing speed or verbal EF was observed. These ndings suggest
the optic radiations are not merely sensory-perceptual relays but also inuence higher-order visual cognition differentially with age.
Keywords: aging, executive function, optic radiations, tractography, white matter connectivity
Correspondence: Kristen M. Kennedy, 1600 Viceroy Dr., Suite 800, Dallas, TX 75235, USA, tel: +1(972)-883-3739, email: kristen.kennedy1@utdallas.edu
Date Received: Dec 10, 2020
Date Accepted: April 12, 2022
DOI: 10.52294/ApertureNeuro.2022.2.ELXU7784
Typical aging is accompanied by the degradation of
structural white matter connections, which are essential
for the propagation of neuronal signals between gray
matter networks. In particular, fractional anisotropy (FA),
a measure of the orientation anisotropy of diffusion,
shows a negative linear association with age and is in-
dicative of deterioration of microstructural organization.
This age-related cortical disconnection is suggested to
contribute to cognitive dysfunction observed in aging,
especially in relation to measures of uid cognition such
as processing speed and executive function (1–5). In sup-
port of this theory, age-related performance variations
in these cognitive domains are shown to be mediated
by age differences in white matter properties (5–13).
This evidence collectively demonstrates that age-related
degradation of white matter connectivity contributes to
poorer cognitive functioning in aging; yet, empirical evi-
dence has largely been limited to the study of major as-
sociation white matter tracts that link heteromodal gray
matter regions supporting higher-order cognitive func-
tioning. Because aging of white matter typically shows
an anterior-to-posterior gradient (13–17), there has been
a greater focus on linking age differences in structural
properties of anterior white matter to age-related varia-
tion in cognition. There has been less research, however,
specically characterizing the extent of aging effects on
posterior relay white matter pathways, such as the optic
radiations (ORs), that support vital sensory-perceptu-
al processing. Moreover, it remains to be determined
whether aging of this white matter inuences not only
performance on speeded tasks, but also more complex
executive functions.
The optic radiations are large white matter bundles
that originate from the lateral geniculate nucleus (LGN)
: 2022, Volume 2 - 2 - CC-BY: © Webb et al.
of the thalamus and extend posteriorly to connect with
the primary visual cortex (18), with a loop through the
temporal cortex (i.e., Meyer’s loop). These radiations are
responsible for the transfer of visual sensory-perceptual
information and are generally thought to be lower-level
input-relay tracts. White matter FA in the optic radia-
tions has been linked to blood-oxygen-level-dependent
(BOLD) activation in the visual cortex (19), and declines
in OR FA can indicate visual impairment in older adults
(20). While the OR facilitates the transmission of visual
information from the retina, this prominent white mat-
ter visual relay system also likely inuences higher-order
cognitive functions. However, the association between
microstructural properties of the OR and their effect on
processing speed or other complex cognitive functions
remains largely understudied, especially in aging. In
children, lower white matter FA in the OR tracts is asso-
ciated with a signicant reduction in processing speed
(21), suggesting that reduced white matter quality of the
OR results in decreased speed at which children process
information and make decisions. Bells et al. (22) further
showed that degraded OR white matter microstructure
is related to neural synchronization in the visual cortex,
which affects cognitive performance in children and ad-
olescents. Additional evidence indicates that individual
differences in younger adults’ performance on a choice
reaction time task are associated with variability of FA in
the OR, yet the sample size of this study was small (23).
Only one study, to our knowledge, has specically linked
white matter properties of the OR with age-related differ-
ences in processing speed. Using principal component
analysis (PCA), Johnson et al. (24) demonstrated that the
OR were indeed sensitive to age-related differences in
FA and that age mediated the association between FA
and perceptual-motor speed. However, the specic re-
lationship between OR white matter and its association
with other higher-order visual and verbal cognitive func-
tions has not been reported in healthy aging adults.
The present study sought to characterize the extent of
age-related effects on FA in the OR in a lifespan sample
of healthy adults. We also aimed to determine whether
aging of OR white matter shows specicity in its associ-
ation with visual versus non-visual, that is, verbal, cog-
nitive performance. We hypothesized that, consistent
with other major white matter tracts, OR white matter FA
would decrease linearly across the age span. Second, in
line with the idea that the OR acts as a visuospatial relay
system supporting lower-level visual and perceptual
functioning, we expected that aging of OR white matter
would be related to basic visual processing speed. Age-
related differences may also exist in the extent to which
FA of the OR is associated with higher-order executive
functions, and notably, may do so differentially across
visual and verbal domains. Specically, aging of the OR
should inuence performance on visual executive func-
tion (EF) tasks to a greater degree than verbal EF tasks.
To address these questions, the present study utilized
diffusion tensor imaging (DTI) and deterministic tractog-
raphy to isolate the OR and characterize both age effects
and age-related differences in cognitive associations in a
large lifespan sample of healthy adults.
Participants included 130 cognitively normal, healthy
adults sampled from across the lifespan ranging in age
from 20 to 94 years (mean age = 47.88±17.36 years;
54 men, 76 women; 78% Caucasian, 11% Black, 9%
Hispanic/Latinx, and 1% Asian), who were recruited from
the Dallas-Fort Worth Metroplex, and compensated
for their time. These 130 participants met the following
study inclusion criteria. Individuals expressing interest in
the study rst underwent an inclusion screening inter-
view conducted by phone that asked individuals about
demographic information (age, date of birth, sex, eth-
nicity, height, weight), previous research study partici-
pation (if participated in a similar cognitive study, they
are excluded due to potential cognitive test exposure/
contamination; if participated in a recent clinical trial for
medications, they are excluded), whether English is a na-
tive language (excluded if learned English after the age
of 6 due to not meeting the norms of the cognitive and
neuropsychological tests and the potential to not per-
form as well due to language issues), handedness (left
handers are excluded due to potential asymmetries in
structural and functional imaging; left handers who were
switched after the age of 4 and remain ambidextrous are
excluded), education (less than high school or GED are
excluded), contraindications for MRI scanning (non-para-
magnetic metal inclusions in body and claustrophobia
are excluded), if pregnant or trying to become pregnant
are excluded (for MRI safety), and head injury with loss of
consciousness (LOC) (excluded if LOC > 5 minutes and/
or if diagnosed or treated for TBI). Participants are also
excluded for diabetes, cancer (if treated with chemother-
apy or radiation), any history of neurological disorders,
any cardiovascular disease or heart problems (except
murmur), psychiatric condition requiring hospitalization,
diagnosed learning disability, vision or hearing problems
(cataracts, glaucoma, colorblindness, macular degen-
eration are excluded, cataracts included if test at nor-
mal visual acuity at study entry, same for hearing aids).
Interview also included open-ended questions about
previous surgeries, other health conditions not previous-
ly mentioned, and if interested in the PET portion of the
study, questions about allergies to pharmaceuticals and
history of liver problems. If interested individuals pass
the phone screening, they are mailed a health ques-
tionnaire packet containing a list of further questions
totaling seven pages, a depression screening tool, and
a more detailed metal screening questionnaire for MRI
: 2022, Volume 2 - 3 - CC-BY: © Webb et al.
(b-value = 1000 s/mm
) with 1 non-diffusion weighted b0
(0 s/mm
), voxel size 2 × 2 × 2.2 mm
(reconstructed to
0.875 × 0.875 × 2 mm
), TR/TE = 5608 ms/51 ms, FOV =
224 × 224, and matrix = 112 × 112. A high-resolution T1-
weighted MPRAGE sequence was also acquired on the
same scanner with the following parameters: 160 sagittal
slices, voxel size 1 × 1 × 1 mm
, ip angle = 12 degrees, TR/
TE = 8.3 ms/3.8 ms, FOV = 256 × 204 × 160, and matrix =
256 × 256.
Diffusion Image Processing and Tractography
DTIPrep v1.2.4 was utilized for processing and quality
control of diffusion images (27) and to identify potential
susceptibility or eddy current artifacts. Gradients with in-
tensity distortions, as well as those of insufcient quality
caused by participant head motion, were detected and
removed from subsequent analyses using the default
thresholds in DTIPrep, and remaining gradients were
co-registered to the non-diffusion-weighted b0 image.
Diffusion directions were adjusted to account for inde-
pendent rotations of any gradient relative to the original
encoding direction (28). DSI studio (29) (software built on
September 26, 2014; http://dsi-studio.labsolver.org) was
used to calculate the diffusion tensors and FA at each
voxel and to conduct deterministic tractography of the
OR. To ensure standardization across the sample, re-
gions of interest and avoidance (ROIs; ROAs) were delin-
eated on either the 1 mm MNI template brain or by using
the individual subject parcellations obtained through
Freesurfer v5.3.0 (30) using the Desikan–Killany atlas (31).
Regions were then warped to individual native subject
diffusion space using a series of non-linear registrations
safety. Individuals passing these inclusion criteria enter
the study provisionally and on rst visit provide written
consent, are tested for visual and hearing acuity, admin-
istered the Mini-Mental State Exam (MMSE) and brachial
blood pressure is measured. Participants scoring < 26 on
the MMSE, testing below corrected speech range fre-
quency, or corrected visual acuity 20/50 or poorer, or have
uncontrolled, unmedicated hypertension are further ex-
cluded from participation in the study. Visual acuity of the
sample included N = 120 participants testing at 20/30 or
better, N = 7 individuals testing at 20/40, and N = 3 in-
dividuals at 20/50. Including visual acuity (unit-weighted
composite z-score of near and far acuity) as a covariate
in nal analyses did not change the pattern of results. All
participants had an MMSE (25) score of at least 26, and a
Center for Epidemiological Studies Depression (CES-D)
(26) score of 16 or lower to exclude participants demon-
strating indications of dementia or depression, respec-
tively. Table 1 reports participant demographics, as well
as cognitive composite scores and mean OR FA, separat-
ed by arbitrary age group; all analyses treated age as a
continuous variable. Table 1 and Figure 1B illustrate that
the age distribution of the sample was roughly rectangu-
lar across the seven decades. The study was approved
by The University of Texas at Dallas and The University of
Texas Southwestern MedicalCenter institutional review
boards, and all participants provided written informed
consent in accord with the Helsinki declaration.
MRI Acquisition Protocol
A 3T Philips Achieva MRI scanner equipped with a 32-chan-
nel head coil was used to acquire diffusion-weighted
images with the following parameters: 65 whole-brain
(WB) axial slices in 30 diffusion-weighted directions
Table 1. Participant demographics and cognitive performance, by arbitrary age grouping
Younger (20–34) Middle (35–54) Older (55–69) Oldest (70–94)
Mean (SD) Mean (SD) Mean (SD) Mean (SD)
Demographics (N) 39 41 33 17
Sex (%F) 56% 54% 61% 71%
Education (years) 15.5 (2.22) 15.3 (2.52) 15.7 (2.56) 16.3 (2.82)
MMSE 29.2 (0.90) 29.3 (0.82) 28.9 (0.80) 28.8 (0.73)
CES-D 4.62 (3.75) 4.22 (4.10) 4.27 (3.71) 3.82 (3.32)
Cognitive Composite Scores
Visual Processing 0.66 (0.73) −0.09 (0.63) −0.26 (0.59) −0.79 (0.55)
Visual EF 0.33 (0.66) 0.12 (0.60) −0.12 (0.67) −0.81 (1.02)
Verbal EF 0.16 (0.74) −0.007 (0.79) −0.05 (0.68) −0.25 (0.83)
Optic Radiations FA 0.498 (0.02) 0.487 (0.02) 0.482 (0.02) 0.480 (0.01)
Note: Age is arbitrarily broken into groups for table description purposes but was treated as a continuous variable in all analyses. Cognitive
composite scores are coded such that higher values represent better performance and/or higher executive functioning. SD, standard devi-
ation; MMSE, Mini-Mental State Exam; CES-D, Center for Epidemiological Studies – Depression Scale; EF, executive function; FA, fractional